Biological Psychology 47 (1998) 1 – 43
Review
The psychological significance of human startle
eyeblink modification: a review
Diane L. Filion a,*, Michael E. Dawson b, Anne M. Schell c
a
Uni6ersity of Kansas Medical Center, Department of Occupational Therapy Edu. 3033 Robinson,
3901 Rainbow Bl6d. Kansas City KS 66160 -7602, USA
b
Uni6ersity of Southern California, Los Angeles, CA, USA
c
Occidental College, Los Angeles, CA, USA
Received 11 September 1996; received in revised form 18 June 1997; accepted 19 June 1997
Abstract
The human startle eyeblink reflex is reliably modified by both cognitive and emotional
processes. This review provides a comprehensive survey of the current literature on human
startle modification and its psychological significance. Issues raised for short lead interval
startle inhibition include its interpretation as a measure of protection of processing, sensorimotor gating and early attentional processing. For long lead interval effects, interpretations
related to attentional and emotional processing are discussed. Also reviewed are clinical
applications to information processing dysfunctions in the schizophrenia spectrum disorders,
and to emotional processing disorders. Finally, an integrative summary that incorporates
most of the cognitive findings is presented and directions for future research are identified
regarding both cognitive and emotional modification of startle. © 1998 Elsevier Science B.V.
Keywords: Human startle eyeblink; Psychological; Sensorimotor gating; Schizophrenia spectrum disorders
* Corresponding author. E-mail: dfilion@kumc.edu
0301-0511/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved.
PII S 0 3 0 1 - 0 5 1 1 ( 9 7 ) 0 0 0 2 0 - 3
2
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
1. Introduction
According to historical accounts of startle research by Hoffman and Ison (1980),
Ison and Hoffman (1983), the fact that reflexes can be reliably modified by
antecedent changes in the sensory environment was first documented in animals in
the mid 1860s by Sechenov (1965). However, it was not until the mid 1970s, when
Graham reviewed literature documenting similar effects with the human startle
eyeblink reflex, that interest in research on human reflex modification began to
show dramatic growth. At that time Graham suggested that the study of startle
modification might ‘‘provide a powerful technique for probing what underlies the
normal processing of information and especially for probing processing characteristics of relatively inaccessible subjects…’’ (Graham, 1975, p. 238). As shown in Fig.
1, in the years since Graham’s suggestion the number of published articles on startle
reflex modification in humans, as indexed by the MedLine and PsycLIT computerized databases, has increased exponentially.
The high level of interest in the startle modification technique is a clear indication
that researchers believe the technique indeed yields valuable information about
human attentional and affective processes. The focus of this review is on the nature
of that information. For what seems to be a rather simplistic measure, interpretations of startle eyeblink modification (SEM) are varied and complex, as SEM has
been reported to index various levels of cognitive and emotional processing.
Much of the complexity associated with the psychological interpretation of SEM
is due to the fact that SEM is affected by many factors including the nature of the
startling and startle modifying stimuli, the temporal configuration of these stimuli,
the instructions given to the research participant, and the characteristics of the
participants themselves. The purpose of this review is to sort through these
methodological and theoretical complexities in order to provide a survey and
integration of the startle modification literature, with the goal of describing the
Fig. 1. Number of published articles on human startle eyeblink modification based on the PsycLIT and
MedLine computerized databases.
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
3
Fig. 2. An illustration of typical startle eyeblink modification paradigms. Panel A shows a baseline
condition in which only a startle-eliciting stimulus is presented. Panel B illustrates a short lead interval
lead stimulation condition resulting in startle inhibition. Panel C illustrates a long lead interval lead
stimulation condition resulting in startle facilitation.
current state of knowledge concerning what can be learned from SEM. There have
been three major reviews written regarding human SEM (Anthony, 1985; Graham,
1975; Lang et al., 1990) as well as an edited book devoted to this topic (Dawson et
al., in press). The purpose of this paper is to review what we have learned about
SEM and from SEM in recent years. This review is focused on the human startle
modification literature. Brief referrals to the extensive literature on startle modification in nonhuman animals will be made for the interested reader where appropriate.
Neurophysiological circuits for both startle and its modification are reasonably well
developed in nonhuman animals, and referrals to this literature will also be given.
2. The basic startle eyeblink modification paradigm
The typical SEM paradigm involves the presentation of a series of trials in which
a startle-eliciting stimulus (e.g. sudden loud noise) is presented in the absence of any
other stimulus (Fig. 2, panel a), intermixed with trials in which the startle-eliciting
stimulus is presented closely following a non-startling stimulus called a ‘lead
stimulus’ (Fig. 2, panels b and c). The interval between the onset of the lead
stimulus and the startle-eliciting stimulus is called the ‘lead interval’. The dependent
measure most commonly reported in human startle research is a change or
percent-change score reflecting the difference in size of the startle eyeblink elicited
under these conditions.
4
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
Procedures for recording and quantifying startle eyeblink amplitude have been
discussed by Anthony (1985), Balaban et al. (1986), Clarkson and Berg (1984) and
Lang et al. (1990). (See Berg and Balaban, in press, for a review). Startle amplitude
inhibition refers to cases in which the startle eyeblink is smaller in the lead
stimulation condition than in the baseline condition (Fig. 2, panel b), startle
amplitude facilitation refers to cases in which the startle reflex is larger in the lead
stimulus condition than in the baseline condition (Fig. 2, panel c), and startle
latency facilitation refers to cases in which the latency of the startle reflex is shorter
in the lead stimulus condition than in the baseline condition.
The nature of the modifying effect of a lead stimulus on startle depends in large
part on the lead interval. Generally speaking, there are two main classes of SEM
effects in humans, (1) modification of amplitude and latency with lead intervals of
up to approximately 500 – 800 ms, which will be referred to as ‘short lead interval
effects’, and (2) modification of amplitude with lead intervals longer than 500–800
ms, which will be referred to as ‘long lead interval effects’. In each of the major
sections of this paper, we will first review the typical reflex modification effects and
their interpretations and then we will discuss the exceptions to these generalities.
3. Short lead interval effects
There are three main types of lead stimulation startle modification effects
observed at short lead intervals. The first is an amplitude facilitation effect which
has been shown to occur reliably for acoustic startle with vibrotactile lead stimuli
at 25, 50, and 100 ms lead intervals (e.g. Blumenthal and Gescheider, 1987;
Blumenthal and Tolomeo, 1989; Flaten and Blumenthal, 1996), for electrically-elicited startle with acoustic lead stimuli at a lead interval of 10 ms (e.g. Boelhouwer
et al., 1991), and has also been observed for acoustic startle with visual and
electrocutaneous lead stimuli at lead intervals of 30 and 60 ms (Graham, 1980). To
date the main focus of research on this amplitude facilitation effect has been on its
physiological significance (e.g. Boelhouwer et al., 1991) and it has received relatively
little empirical attention or discussion in terms of its psychological significance.
The second short lead interval startle modification effect is a latency facilitation
that is observed at lead intervals of approximately 100 ms or less, regardless of the
modalities of the lead and startle stimuli (e.g. Graham, 1975; Graham and Murray,
1977; Braff et al., 1978; Blumenthal and Gescheider, 1987; Blumenthal and
Tolomeo, 1989). The psychological significance of this latency facilitation effect has
also received relatively little attention, the effect has been reported inconsistently
(many reports do not include the latency measure at all), and is often reported as
merely co-occurring with amplitude inhibition effects (discussed below). However,
there is considerable evidence suggesting that latency facilitation and amplitude
inhibition are independent processes that show a different developmental course
(e.g. Anthony and Graham, 1985; Ornitz et al., 1986, 1991), are affected differently
by various startle and lead stimulus parameters (e.g. Graham and Murray, 1977;
Silverstein et al., 1981; Blumenthal and Levey, 1989), and psychopathological states
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
5
(e.g. Braff et al., 1992). Thus latency facilitation effects may deserve greater
attention in future research, particularly in future research involving clinical populations.
The main focus of this section is the third and most robust short lead interval
effect, inhibition of startle amplitude, an effect also referred to in the literature as
prepulse inhibition (PPI). Excluding the conditions described above known to
produce amplitude facilitation, amplitude inhibition is produced by a wide range of
lead stimuli and occurs within a lead interval range of approximately 30–500 ms.
The startle amplitude inhibition effect is quite robust, typically in the range of
50– 80% inhibition, and is quite reliable, occurring in 90–100% of normal adult
participants who show reliable eyeblink reflexes. Startle inhibition can be produced
by visual, acoustic, olfactory and vibrotactile lead stimuli, and their inhibitory
effect is seen even when the modalities of the lead and startle stimuli differ.
Moreover, an increase or a decrease in stimulus energy can also serve as an effective
lead stimulus. Startle inhibition has been obtained in infants (e.g. Balaban et al.,
1989; Graham et al., 1981), though the inhibitory processes do not appear to be
fully developed until around the age of eight (Ornitz et al., 1986, 1991). (See Ornitz,
in press, for a review). Startle inhibition is well documented in nonhuman animals
(see review by Ison and Hoffman, 1983). In nonhuman animals the neural substrates of both the acoustic startle reflex (e.g. Davis et al., 1982; Koch et al., 1992;
Lee et al., 1996; Davis et al., in press) and its inhibition at short lead intervals (e.g.
Geyer et al., 1990; Leitner and Cohen, 1985; Swerdlow et al., 1992) are also well
documented (for current reviews of this work see Dawson et al., 1997; Swerdlow
and Geyer, in press).
As noted by Blumenthal (in press), ‘‘Short lead interval modification of startle is
a response measure that has several obvious advantages, including: (1) availability
of animal models, leading to an understanding of the neurological mechanisms
underlying the effect; (2) availability of developmental models; (3) minimal compliance and motivation required of the subject; (4) sensitivity to manipulations of the
sensory, cognitive, social, and pharmacological environment; (5) an effect size that
is great enough that even rather large changes in methodology cannot obscure this
effect; (6) functional significance in the life of the organism. What more can we ask
of a response measure?’’
In the following sections we will discuss the current views of the psychological
interpretation and functional significance of this inhibitory effect in humans.
3.1. Protection of processing
In the first review of startle inhibition effects in humans, Graham (1975) raised
the possibility that the inhibition may reflect ‘‘a wired-in negative feedback which
reduces the distraction produced by reflexes such as startle, and thus protects what
has been called pre-attentive stimulus processing’’ (p. 246). Graham’s view is that
the onset of low-intensity changes in sensory stimulation evokes a ‘transient
detecting reaction’ that automatically triggers a sensory-gating mechanism which
momentarily prevents or attenuates extraneous reactions such as startle until the
6
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
perceptual analysis of the lead stimulus is completed. According to Graham (1975,
1979, 1992) two parallel processes occur when a stimulus is perceived: one is the
encoding and perceptual analysis of the stimulus, and the other is a protective
process which attenuates all subsequent stimulation until the perceptual/encoding
processes have been completed. Two primary lines of evidence supporting the
protection-of-processing hypothesis are discussed below.
3.1.1. Startle inhibition and startle stimulus perception
The first line of support consists of two studies that examined the perceived
intensity of the startle stimulus. These studies tested the hypothesis that the startle
stimulus would be perceived as less intense if its impact on the system were reduced
by the lead stimulus. In these studies, the primary dependent variables were
measures of startle response magnitude elicited under baseline and lead stimulus
conditions and participant ratings of the intensity of the startle-eliciting stimulus
under the same conditions. Using an 80 dB tone as a lead stimulus and a tap to the
forehead (glabellar tap) as the startle-eliciting stimulus, Cohen et al. (1981) found
that the presence of the lead stimulus decreased both the size of the eyeblink elicited
by the tap as well as the estimated intensity of the tap itself. The same pattern of
results was obtained by Perlstein et al. (1993), using a 75 dB tone as a lead stimulus
and a 110 dB tone as the startle-eliciting stimulus. Although these studies appear to
support the protection-of-processing hypothesis, Anthony (1985) has suggested that
in the Cohen et al. and by extension, the Perlstein et al. studies, participants’
estimates of startle stimulus intensity may have been biased by the perception of
their own reduced startle response magnitude. Blumenthal et al. (1996), however,
recently demonstrated that with high intensity startle-eliciting stimuli (i.e. 100 dB or
greater) both startle response magnitude and ratings of startle stimulus intensity are
reduced by lead stimulation, but the amount of reduction between these measures
is not correlated across participants. In any event, it may be as important to protect
processing from disruption by efferent motor responses to the startle stimulus as it
is to protect it from disruption by afferent sensory input itself.
3.1.2. Startle inhibition and lead stimulus perception
The second line of investigation testing the protection of processing hypothesis
consists of four studies that have examined the impact of a startle-eliciting stimulus
on participants’ perceptions of the lead stimulus. If startle inhibition serves to
protect the perceptual processing of the lead stimulus, then perception of the lead
stimulus should be more accurate when it is effective in producing startle inhibition.
In each of two studies, (Filion and Ciranni, 1994; Perlstein et al., 1993, experiment 2), participants were presented with three stimulus conditions relevant to the
present discussion: a lead stimulus alone, a startling stimulus alone, and the lead
stimulus followed by the startling stimulus at a lead interval of 500 ms (Perlstein et
al., 1993) or 120 ms (Filion and Ciranni, 1994). The primary dependent measures
were startle-eyeblink amplitude and participant ratings of lead stimulus intensity
recorded under the paired and non-paired conditions. In terms of startle responding, both studies confirmed the basic inhibition effect in the lead stimulus condition.
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
7
In terms of the lead stimulus intensity ratings, the lead stimulus in the paired
condition was rated as slightly, but significantly, louder than it was in the
non-paired condition, indicating that in the paired condition, the startling stimulus
interfered with the perceptual analysis of the lead stimulus. This is a well-documented phenomenon known as loudness assimilation. This result suggests that
inhibition of the startle response in the lead stimulation condition did not result in
perfect protection of the perceptual analysis of the lead stimulus. However, as a
further test of the relationship between startle inhibition and lead stimulus perception, Filion and Ciranni (1994) also computed the correlation between the amount
of startle inhibition produced by the lead stimulus and the amount of interference
in lead stimulus perception produced by the startle stimulus. In support of the
protection-of-processing hypothesis, a significant correlation was obtained indicating that participants showing greater startle inhibition had a more accurate
perception of the lead stimulus. Similarly, using the same paradigm, Perlstein,
Fiorito, Simons, and Graham (1989) discussed in Graham (1992) found that college
students who scored abnormally high on a perceptual aberration scale exhibited
significantly less startle inhibition as well as significantly greater startle-produced
interference with lead stimulus perceptions than did a normal control group.
Finally, following the same logic but taking a slightly different strategy, Norris
and Blumenthal (1995) required participants to indicate after each trial whether a
high-pitched lead stimulus, a low-pitched lead stimulus, or no lead stimulus had
been presented. Because the tone pitches were difficult to discriminate, these
investigators were able to use the number of hits and misses for the target lead
stimulus as a measure of the accuracy of lead stimulus perception. Consistent with
the protection-of processing hypothesis, results revealed that greater startle inhibition was produced on trials in which the lead stimulus was correctly identified than
on trials in which the participant’s response was incorrect.
3.1.3. Protection of processing and sensorimotor gating
Taking a view quite similar to the protection-of-processing hypothesis, Braff,
Geyer, and colleagues (e.g. Braff and Geyer, 1990) have suggested that startle
inhibition may serve as an operational measure of sensorimotor gating, ‘‘reflecting
the ability to effectively buffer or screen out the potentially chaotic flow of
information and sensory stimuli’’ (Cadenhead et al., 1993, p. 1862). In their view,
startle inhibition indexes a basic inhibitory process which regulates sensory input to
the brain and allows the early stages of information processing to occur without
disruption. What sets this view apart from the protection-of-processing view is the
suggestion that startle inhibition reflects a general ability to inhibit external stimuli
(auditory, visual, tactile, etc.) as well as internal stimuli such as thoughts and
impulses (e.g. Geyer et al., 1990).
In support of the sensorimotor gating view, there is a rapidly growing literature
documenting startle inhibition deficits in clinical populations characterized by
inhibitory deficits and an inability to regulate internal stimulation (to be reviewed
in a later section). To date, startle inhibition deficits have been reported in
schizophrenia patients (Braff et al., 1978, 1992; Dawson et al., 1993; Grillon et al.,
8
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
1992), schizotypal patients (Cadenhead et al., 1993), college students scoring high
on psychosis-proneness scales (Perlstein et al., 1989; Schell et al., 1995; Simons and
Giardina, 1992; Swerdlow et al., 1995a; but see Blumenthal and Creps, 1994 and
others to be reviewed later for exceptions), obsessive-compulsive disorder patients
(Swerdlow et al., 1993), individuals with Huntington’s Disease (Swerdlow et al.,
1995b), and children with attention-deficit disorder (Anthony, 1990) and nocturnal
enuresis (Ornitz et al., 1992). In addition, Perry and Braff (1994) reported a
significant positive relationship between startle inhibition deficits and independent
measures of thought disorder in a population of individuals with schizophrenia.
Taken together, these findings suggest that startle inhibition may indeed index an
inhibitory ability that generalizes beyond specific time-locked inhibitory effects
following the onset of external stimuli.
3.2. Protection of processing: summary
The studies above are generally supportive of the protection-of-processing hypothesis and suggest that startle inhibition may be mediated by a partial sensory
blockade of the startle stimulus. One direction for future tests of the protection-ofprocessing hypothesis may be to design studies that assess more independent
measures of startle stimulus impact. For example, Foss et al. (1989) found that
presentation of a weak lead stimulus 100 ms prior to a startle-eliciting stimulus
significantly reduced startle-produced errors in rifle-aiming, suggesting that lead
stimulation does reduce the impact of the startle-eliciting stimulus on motor output.
Further work is needed to determine whether startle responses can be shown to
disrupt performance on non-motor tasks, and whether presentation of a lead
stimulus can reduce such disruption. Finally, although it is consistent with the
sensory gating view that several clinical populations characterized by reduced
inhibition also show SEM deficits, further research is needed to determine whether
these deficits signify a common cognitive impairment, share a common underlying
neurophysiology, or both. Although the bulk of the research described above
supports the protection-of-processing hypothesis, many questions remain as to the
nature of the protective process.
3.3. Early attentional processing
One question raised by the protection-of-processing and sensorimotor gating
views concerns whether the protection of lead stimulus processing occurs automatically or whether it is dependent upon controlled attentional processing.
3.3.1. E6idence for automatic processes
Recall that Graham (1975, 1979) originally suggested that short lead interval
startle inhibition might reflect the protection of pre-attentive processing, implying
that the inhibitory effect should be attention invariant. This suggestion was based
on the findings that startle inhibition is observed in nonhuman animals, and in fact
in decorticate rats (Ison et al., 1991), as well as in infants (Graham et al., 1981),
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
9
sleeping human adults (Silverstein et al., 1980), and that it is observed on the first
presentation of a lead stimulus-startle stimulus pairing (Graham, 1975). Moreover,
the fact that pre-habituation of the lead stimulus does not affect startle inhibition
in either nonhuman animals (Wu et al., 1984) or humans (Wynn et al., 1996)
suggests that this inhibition reflects the action of automatic processes. Clearly,
evidence indicates that short lead interval startle inhibition is at least partially an
automatic effect. However, this does not preclude the possibility that this startle
inhibition might be modulated by controlled attentional processes. The following
section describes several studies designed to examine the modulatory effects of
attention on startle inhibition.
3.3.2. E6idence for attentional processes
The possibility that attention may have a modulatory influence on SEM was
raised initially by studies documenting an increase in startle inhibition in conditions
in which a warning stimulus predicted the presentation of the lead stimulus (e.g.
DelPezzo and Hoffman, 1980; Hackley and Graham, 1983; Ison and Ashkenazi,
1980). In the DelPezzo and Hoffman (1980) studies, for example, the eyeblink
eliciting stimulus was a tap to the forehead and the lead stimulus was a light
presented at a lead interval of 150 ms that could appear in one of several possible
locations on a grid located directly in front of the participant. Participants were
instructed to maintain their gaze at a position in the center of the grid. In the first
experiment, participants were warned on half of the trials as to the location where
the lead stimulus would be presented, and were given no information about the lead
stimulus on the remaining trials. The result of interest is that startle inhibition was
greater on trials in which the participants knew the location of the upcoming lead
stimulus than when they did not. This finding suggests that directing attention
toward the location of a visual lead stimulus increases the amount of inhibition
produced by that lead stimulus. The second experiment tested the hypothesis that
instructing participants to direct their attention away from a visual lead stimulus
would decrease the amount of inhibition produced by that lead stimulus. In this
experiment, the lead stimulus consisted of a light that was always presented at the
center position of a grid located directly in front of the participant. While
maintaining visual fixation on this center location, participants were instructed
prior to half of the light presentations to attend to the central location, and prior
to the other half of the trials to attend to grid location 40° to the left or right of
the center location. The results of this procedure revealed that the presentation of
the visual lead stimulus produced the expected startle inhibition, and that the
inhibition was greater when attention was directed to the location of the lead
stimulus than when attention was directed away from the lead stimulus.
Consistent with this finding, McDowd et al. (1993) found greater startle inhibition at a lead interval of 120 ms during a trial block in which participants were
required to press a response key whenever they detected a lead stimulus than during
a trial block in which there were no task instructions. Taken together these studies
suggest that, contrary to what would be expected with a purely automatic process,
directing attention toward a lead stimulus increases the amount of startle inhibition
10
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
it produces. However, the attention effects may be attributed to generalized
attentional/arousal processes rather than selective attentional processes because
attention was manipulated across blocks of trials (Harbin and Berg, 1986).
In an attempt to address the question of whether the attentional effects on startle
inhibition are general or selective, Acocella and Blumenthal (1990) employed an
intense white noise as the startle eliciting stimulus and a low pitched tone, a high
tone, and a less intense white noise as three unique lead stimuli. On one block of
trials participants were instructed to attend to the lead stimuli and to press a
response key whenever a particular stimulus was presented (e.g. press whenever you
hear a low tone). On another block of trials, participants were instructed to ignore
all lead stimuli. Results indicated that the three lead stimuli all produced startle
inhibition, and that inhibition was greater when participants were attending to the
lead stimuli than when the lead stimuli were ignored. Within the attention block of
trials, however, there were no differences observed between the amount of inhibition produced by the lead stimulus that signalled a motor response and the other
two lead stimuli. These results suggest a generalized attentional/arousal effect rather
than a selective attention effect.
Hackley and Graham (1987, experiment 2) also examined attentional effects on
startle inhibition, but employed a task that did not require a motor response from
the participant. Instead, startle inhibition was examined under conditions in which
to-be-attended and to-be-ignored lead stimuli were presented within the same trial
block in an unpredictable sequence. In this experiment, a warning signal was
presented, followed by the presentation of a tone to either to the left ear, the right
ear, or to both ears simultaneously. Participants were required to judge the
durations of the tones presented to one spatial location while ignoring tones
presented to the other locations. A subset of these to-be-attended and to-be-ignored
tones served as lead stimuli for a startle-eliciting air puff. The results indicated that
the lead stimuli reliably produced startle inhibition, and the attended lead stimuli
produced greater startle inhibition than the ignored lead stimuli when the discrimination was between lateralized stimuli.
Expanding upon these results, Filion et al. (1993) conducted an experiment in
which participants were presented an unpredictable intermixed series of tones of
two different pitches and were instructed to count the number of longer than usual
occurrences (i.e. 7 rather than 5 s) of one tone pitch and to simply ignore the tones
of a different pitch. This differential attention paradigm was originally devised and
employed to study attentional processes associated with the orienting response
(Dawson et al., 1989; Filion et al., 1991). Specifically, it requires participants to
perform a series of cognitive operations beginning with the onset of a tone: (1) to
first pre-attentively detect and discriminate the to-be-attended tone from the
to-be-ignored tone; (2) to then allocate additional controlled processing resources
necessary to confirm the identity of the to-be-attended tone and begin the duration
judgment task; and (3) finally, to sustain selective attention to the to-be-attended
tone throughout its duration in order to determine its length and perform the
required timing task (Dawson et al., 1997).
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
11
In the Filion et al. (1993) study, the two tones served as lead stimuli for a
startle-eliciting noise burst presented at short lead intervals of 60, 120, and 240 ms.
The results, re-presented in Fig. 3, revealed significant startle inhibition at each of
these short lead intervals, and that for the 120 ms lead interval, the attended lead
stimulus produced significantly greater startle inhibition than the ignored lead
stimulus. These basic results have since been replicated and extended in six
independent samples: unselected college students (Filion et al., 1994; Jennings et al.,
1996; Seljos et al., 1994), college students selected to serve as normal controls for
high-risk populations (Schell et al., 1995), and non-college students selected to serve
as normal controls demographically matched to schizophrenia outpatients (Dawson
et al., 1993; Hazlett et al., 1995).
In another recent study relevant to the discussion of attentional effects on SEM,
Hazlett et al. (1993) presented startle stimuli at short lead intervals following the
onsets of digits presented in a memory load version of a continuous performance
task. In this task, participants viewed a series of rapidly presented digits on a
computer screen and were asked to press a response key whenever they detected the
digit ‘7’ immediately following the digit ‘3’. A startle eliciting noise burst was
presented at lead intervals of 120 and 240 ms following the onset of a subset of the
digits assumed to be associated with high attentional processing (e.g. the target digit
‘3’), as well as digits assumed to be associated with relatively low attentional
processing (e.g. non-target digits temporally distant from both 3 and 7 s). Results
of this procedure revealed significantly greater startle inhibition at the 120 ms lead
interval following the target lead stimuli than non-target lead stimuli. Consistent
with these results for target and non-target stimuli, Bradley et al. (1996) found
greater startle inhibition at a lead interval of 300 ms by lead stimuli consisting of
positively and negatively valenced pictures than by neutral pictures, which were
presumably less arousing, interesting, and/or attention-demanding.
Fig. 3. Short lead interval startle eyeblink modification results from an active attention paradigm
(adapted from Filion et al., 1993).
12
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
Together, these results demonstrate that attention can modulate startle inhibition
at short lead intervals in a variety of paradigms, and that the effects can be
observed cross-modally, with visual lead stimuli and acoustic startle-eliciting stimuli. It should be noted, however, that Vanman et al. (1996, experiment 1) observed
no difference in inhibition at a lead interval of 250 ms between affective pictorial
lead stimuli that were to be attended compared to affective pictures that were to be
ignored. Although this finding appears inconsistent with the above studies, an
interpretation suggested by Vanman et al. (1996) is that the affective/arousing
nature of the pictures used in this experiment may have overwhelmed the effect of
the instructional attentional manipulation.
The studies just described suggest that attentional manipulations may have a
selective, stimulus specific effect on the startle inhibition produced by a to-be-attended lead stimulus, although a nonspecific arousal/activational effect may have
contributed to the SEM observed to the to-be-ignored lead stimulus as well. To
date, however, only one study has examined the relative contributions of attentional
effects that are due to non-specific increases in arousal/activation versus effects that
are due to selective stimulus processing. Jennings et al. (1996) tested participants in
one group with the tone-length judgment task of Filion et al. (1993) and tested
participants in a second group with identical stimuli but with no task to perform.
Using this design, selective attentional effects can be examined by comparing the
amount of inhibition produced by the to-be-attended and to-be-ignored lead stimuli
in the selective attention group, whereas non-specific activational effects can be
examined by comparing the startle inhibition produced by the to-be-ignored lead
stimulus in the selective attention group to the amount of inhibition exhibited by
the group given no explicit instructions regarding the lead stimuli and no task to
perform.
In this experiment, lead stimuli were high and low pitched tones which preceded
a startle-eliciting noise burst at a short lead interval of 120 ms. Results for the
active attention group revealed that the to-be-attended lead stimulus produced
greater startle inhibition than the to-be-ignored lead stimulus. In addition, greater
inhibition was produced by the to-be-attended prepulse in the active attention
group than by the comparable lead stimulus in the passive attention group, whereas
the amount of inhibition produced by the to-be-ignored lead stimulus in the active
attention group did not differ from the inhibition produced by the comparable lead
stimulus in the passive attention group. These results suggest that the attentional
instructions had a selective effect in that they increased startle inhibition to the
to-be-attended lead stimulus, whereas there was no evidence of a non-specific
arousal effect because the inhibition produced by the to-be-ignored lead stimulus in
the attention group and the comparable lead stimulus in the no-task group did not
differ.
3.4. Early attentional effects: summary and conclusions
Taken together, the studies just reviewed suggest that startle inhibition at short
lead intervals may reflect not only the automatic protective and/or sensorimotor
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
13
gating system, but that it may also index the early stages of controlled attentional
processing. As discussed further in the concluding section of this paper, the
automatic and controlled processes appear to follow different time courses and may
be separable by studying the effects of the task-imposed attentional demands of the
lead stimulus. That is, findings from several studies indicate that attentional
modulation of startle inhibition occurs only at and after approximately 120 ms, and
not during the earlier portion of the time window of short lead interval inhibition,
suggesting that controlled attentional processes develop more slowly. However, it is
also important to note that in several of these studies, the attentional enhancement
of startle inhibition occurred at a 120 ms lead interval, but not at a 240 ms lead
interval, suggesting that the attentional processes underlying this effect have an
extremely short duration or that the attentional processes may interact with the
processes underlying the transition from short lead interval startle inhibition to long
lead interval startle facilitation. Thus, further research is needed to determine which
stage(s) or type(s) of attentional processing are indexed, as well as the relationship
between attentional effects and more non-specific arousal effects.
3.5. Relationship to other measures
3.5.1. Beha6ioral measures
As reviewed by Filion et al. (in press), another strategy for assessing the
psychological significance of short lead interval startle inhibition is to examine its
relationship to behavioral measures of inhibition such as performance on backward
masking, negative priming, and Wisconsin Card Sorting Tasks. Although attempts
to correlate these behavioral measures with startle inhibition on a within-participant basis have yielded inconsistent results, psychopathological populations (particularly those with schizophrenia spectrum disorders) who exhibit reduced startle
inhibition have been shown to exhibit consistent patterns of deficits in these tasks.
These results suggest that the inhibitory processes reflected by these measures may
be different, occur at different stages of information processing, and/or have
different neural substrates. Further research is needed to determine more precisely
the nature and generality of the inhibitory processes reflected by short lead interval
startle inhibition.
3.5.2. E6ent-related potentials
Another line of investigation possibly related to the significance of the short lead
interval startle inhibition effect involves the examination of event related potentials
in lead stimulation paradigms. Although few studies have examined evoked potentials in what would be considered a typical startle inhibition paradigm (e.g. a low
intensity lead stimulus followed by a high intensity startle-eliciting stimulus), there
is a large literature in which the P50 component of the cortical evoked potential has
been recorded during presentation of pairs of moderately loud auditory clicks
separated by approximately 500 ms. The P50 amplitude elicited by the second test
click is normally suppressed compared to the P50 elicited by the first member of the
pair. This P50 suppression effect is interpreted as indexing an inhibitory sensory
14
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
gating process similar to the processes thought to be indexed by startle inhibition at
short lead intervals (Freedman et al., 1987).
In one attempt to examine the relationship between short lead interval startle
inhibition and P50 suppression, Schwarzkopf et al. (1993) tested the same participants using the typical protocols for each measure. The two test protocols, with a
lead interval of 100 ms for startle inhibition and 500 ms for P50 suppression, were
separated by 20 min. Only a modest trend for a positive association was found
between P50 suppression and startle inhibition. Moreover, the fact that the optimal
lead interval for P50 suppression is longer than the optimal interval for startle
inhibition further calls into question whether the two measures are tapping the
same inhibitory processes. In order to obtain measures of startle inhibition and
modification of evoked potential concurrently within exactly the same paradigm,
Perlstein et al. (1993) recorded evoked potentials and startle eyeblinks simultaneously under four conditions: low intensity tone alone (75 dB(A), 40 ms 800 Hz
tone); high intensity tone alone (110 dB(A), 40 ms 800 Hz tone); low intensity-low
intensity pair separated by a fixed lead interval; and low intensity high-intensity
pair separated by the same lead interval. Significant inhibition of the startle
eyeblink occurred at both a 120 ms (experiment 1) as well as a 500 ms (experiment
2) lead interval in both of the paired-stimulus conditions compared to the single
stimulus controls. The results revealed that the P30 component of the evoked
potential was unaffected by stimulus pairing, but that the P50, N100, and P200
components were significantly smaller in response to the second tone in the paired
stimulus conditions compared to the single tone baseline. Extending this work,
Sugawara et al. (1994) examined lead stimulation effects on the P300 by comparing
the P300 elicited by a startle stimulus alone (104 dB SPL, 50 ms, white noise), and
by lead stimulus (75 dB SPL, 25 ms, 1000 Hz tone)-startle stimulus pairings at lead
intervals of 60 and 120 ms. Startle inhibition as well as a significant reduction in the
P300 occurred to the startle-eliciting stimulus presented in the lead stimulation
conditions compared to the response to the startle-eliciting stimulus alone. These
results demonstrate that conditions producing startle inhibition also produced
reductions in evoked potential components occurring in the mid to long latency
range (50 – 300 ms). These reductions would generally be expected if sensorimotor
gating processes are reducing the effective intensity of the startle eliciting stimulus.
(For a further analysis of parallels and differences between lead stimulation effects
on startle and on ERP components, see Ford and Roth, in press).
4. Long lead interval effects
In contrast to the unidirectional inhibitory effects discussed above, at lead
intervals greater than approximately 800 ms, startle amplitude can exhibit either
inhibitory and facilitatory effects, depending on stimulus conditions and participant
instructions (see reviews by Anthony, 1985; Graham, 1992; Lang, 1995; Putnam,
1990; Putnam, in press). In addition, in contrast to the homogeneous interpretation
of short lead interval effects involving the protection-of-processing/sensorimotor
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
15
gating frameworks, at long lead intervals there are two main classes of modification
effects, modification by attentional processes and modification by emotional processes, and each is interpreted within a different theoretical framework and apparently involves different neural pathways. In the following sections we first review
each class of modification effect and its interpretation, and then review a small
literature consisting of experiments designed to examine the relative contributions
of attentional and emotional processes to SEM within the same participants and
within the same paradigm.
4.1. Modulation by attention
Research with nonhuman animals has demonstrated that the amplitude of the
startle reflex is facilitated when startle is elicited after a relatively long lead interval
(e.g. 2000 ms) if the lead stimulus is sustained throughout the lead interval
(Hoffman and Wible, 1969; see Graham, 1975, pp. 243–246, for a review). Graham
(1975) suggested that the mechanism underlying startle facilitation with continuous
lead stimulation may be ‘‘a classical activation effect mediated by the reticular
activating system’’. Graham et al. (1975) replicated these effects with human startle
eyeblink and found, unlike the earlier findings with rats, that startle facilitation also
occurred with discrete lead stimuli (those that offset before the end of the lead
interval) as well as with continuous lead stimuli. The human participants also
showed strong and persistent heart rate orienting responses during the lead interval,
particularly when the onset of the startle-eliciting stimulus was unpredictable,
suggesting that orienting and attention were playing a role in SEM in addition to
the effects of ‘classical activation’. These observations led to a systematic series of
investigations and hypothesizing about attentional factors influencing human SEM
at long lead intervals.
Early investigations of attentional influences on SEM at long lead intervals
focused primarily on modality-specific selective attention effects. Studies have
shown that at long lead intervals startle is facilitated when participants are
instructed to focus their attention on the startle-eliciting stimulus or on the same
sensory modality as the startle-eliciting stimulus (Bohlin and Graham, 1977; Bohlin
et al., 1981; Hackley and Graham, 1983). When instructions direct attention toward
a modality different from that of the startle-eliciting stimulus, startle is inhibited
(Anthony and Putnam, 1985; Putnam, 1990; Silverstein et al., 1981). These effects
are interpreted as being due to a process of selective sensory pathway enhancement
whereby paying attention to a sensory modality primes the sensory pathway for
that modality, thereby facilitating sensory processing of all input in that modality,
and simultaneously inhibiting sensory processing of other modalities. Thus, attention to the startle eliciting stimulus is viewed as turning up the gain in certain
anatomical structures which overlap those involved in the elicitation of the startle
blink (Anthony, 1985).
One implication of this view is that SEM effects, both facilitatory and inhibitory,
should be enhanced in situations in which the degree of attention as well as the
direction of attention is manipulated. This notion was tested explicitly in a series of
16
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
Fig. 4. Startle responses to auditory eliciting stimuli presented during the 6000 ms warning interval of
an anticipation task (adapted from Putnam, 1990).
studies by Putnam and colleagues (see Putnam, 1990 for a review), who presented
startle-eliciting stimuli during the warning interval of a speeded simple reaction
time task. Since this type of task requires participants to attend to the modality of
the imperative stimulus and to ‘pay closer attention’ as the end of the warning
interval approaches, this paradigm allows examination of the effects of both the
direction and degree of attention on SEM. Consistent with the view that long lead
interval SEM is sensitive to both the direction and degree of attention, not only
were startle responses facilitated when the modalities of the startle-eliciting stimulus
and the expected imperative stimulus matched, and were inhibited when the
modalities were different, these attentional effects were stronger for startle-eliciting
stimuli presented later compared to earlier in the warning interval. These findings
are illustrated in Fig. 4.
In another set of studies examining attentional influences on long lead interval
SEM, Anthony and Graham (1983, 1985) presented one group of participants with
a set of visual lead stimuli, half consisting of light-coloured blank slides (considered
dull), and half consisting of slides of human faces (considered interesting). A second
group of participants received auditory lead stimuli consisting of either tones
(considered dull), or music-box melodies (considered interesting). The startle-eliciting stimulus for both groups, presented after a lead interval of 4000 ms, was
auditory (a noise burst) on half of the trials, and was visual (a light flash) during
the remaining trials. The results of this procedure with infants indicated that for
each startle stimulus modality, startle responses were larger when the lead and
startle-eliciting stimuli matched than when they were different. In addition, attention effects were observed within modalities such that when the modalities of the
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
17
lead and startle-eliciting stimuli matched, startle responses were larger when the
lead stimulus was interesting than when it was dull. This same pattern of results
was obtained with adult participants except that the attentional effects were
observed in the latency of the startle responses rather than the amplitude. Anthony
and Graham (1985) argue that the modality specific effects suggest that SEM is due
to modification in specific afferent pathways rather than to modification in a final
common motor pathway. Findings consistent with this viewpoint were reported by
Simons and Zelson (1985), who found that startle responses elicited by auditory
stimuli were smaller when presented during interesting compared to dull visual lead
stimuli, and by Zelson and Simons (1986), who found that performance of a visual
vigilance task resulted in significant inhibition of the acoustic startle response, and
that a difficult vigilance task produced greater inhibition than a less difficult
vigilance task.
Also consistent with these findings, Filion et al. (1993, 1994), in the studies
described earlier, presented participants with a series of intermixed high and low
pitched tones which varied between 5 and 7 s in duration and instructed them to
keep a count of the number of seven second tones of one pitch and to ignore the
tones of the other pitch. Results revealed that both to be-attended and to-be-ignored tones produced startle facilitation at a lead interval of 2000 ms, and that
there was greater facilitation following the attended than the ignored tones. This
pattern of findings using this paradigm has been replicated several times in our
laboratory (Dawson et al., 1993; Jennings et al., 1996; Schell et al., 1995; Seljos et
al., 1994). The greater startle facilitation following the attended than the ignored
lead stimulus has been interpreted as reflecting greater allocation of attentional
resources to the attended tone.
This interpretation was supported by Jennings et al. (1996) in which the same
task was used and startle-eliciting stimuli were presented at lead intervals of 2000,
4500, and 6000 ms. This tone length judging task should result in increasing
attentional demands across the lead stimulus window (that is, participants should
pay closer attention toward the end of the tone when the critical length judgment
is to be made). Therefore, Jennings et al. (1996) predicted that if startle facilitation
reflects the allocation of attentional resources, then startle facilitation should show
a commensurate increase across the lead intervals. An additional feature of this
study was that for half of the trials involving the 6000 ms lead interval, the lead
stimulus duration was 5 s and for half it was 7 s. This allowed an additional test of
the attentional resource hypothesis, by comparing facilitatory effects when the
startle-eliciting stimulus is presented at a time point when attention was assumed to
be high and at a time point when attentional processing was assumed to be
completed. Fig. 5 shows the long lead interval results from this experiment.
Consistent with the attentional resource hypothesis, the results revealed that both
the attended and ignored tones produced startle facilitation with greater facilitation
produced by the attended tone, and that during the attended tone greater facilitation was observed at the 4500 ms lead interval than the 2000 ms lead interval. There
was greater facilitation at the 6000 ms lead interval during the 7 s lead stimuli than
at the 2000 ms lead interval, but there was no facilitation at the 6000 ms lead
18
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
interval during the 5 s lead stimuli. Thus, the results of Jennings et al. (1996)
indicate that the amount of startle facilitation at long lead intervals is a function of
the degree of attentional processing.
In order to investigate the generality of these effects, two studies have employed
the length judging task to determine whether the long lead interval attentional
modulation effects are modality specific (Böhmelt et al., 1996; Lipp et al., 1997
experiment 2; Lipp and Siddle, in press). Lipp et al. examined the effects of lead
stimulus modality as a between-groups factor by presenting an acoustic startle
stimulus at lead intervals of 3500 or 4500 ms during to-be-attended and to-be-ignored lead stimuli that were acoustic for one group of participants and visual for
a second group. The results revealed greater startle facilitation during to-be-attended than to-be-ignored lead stimuli in both groups. Böhmelt et al. (1996)
examined the effects of lead stimulus modality as a within groups factor by
presenting participants with two classes of auditory lead stimuli (high pitch tones
and low pitch tones), and two classes of visual stimuli (horizontal and vertical bars).
Participants were instructed to count the number of occurrences of one of the
auditory and one of the visual stimuli that had a longer than usual duration (e.g.
count the number of high tones and vertical bars that are of longer than usual
duration). These stimuli served as lead stimuli for a startle eliciting noise burst
presented at a long lead interval of 4500 ms. The results revealed that the
to-be-attended tone produced greater startle facilitation than the to-be-ignored tone
and also that the to-be-attended visual stimulus produced greater startle facilitation
than the to-be-ignored visual stimulus. Thus, the findings of Lipp et al. (1997) and
Böhmelt et al. (1996) suggest an attentional effect on long lead interval SEM that
is independent of stimulus modality.
Fig. 5. Long lead interval startle eyeblink modification results from an active attention paradigm
(adapted from Jennings et al., 1996).
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
19
Together, the studies reviewed in this section demonstrate that attentional
processes can have bi-directional effects on the startle reflex at long lead intervals.
Further research is necessary to determine whether the attentional processes have
effects that are modality specific, modality independent, or both, depending upon
the nature of the attentional task. A related question is whether this modulation of
startle occurs in afferent pathways, efferent pathways, or both.
4.2. Modulation by emotion
Another important type of SEM effect observed at long lead intervals is that seen
in the processing of affective stimuli. In nonhuman animals, the whole body startle
response is reliably facilitated by presenting the startle-eliciting stimulus at a long
lead interval following a stimulus that has previously been predictive of an aversive
event such as a shock (e.g. Brown et al., 1951; Davis and Astraehan, 1978), a
phenomenon termed the fear-potentiated startle effect. This effect has been observed over a wide range of lead intervals but appears to be maximal when the lead
interval matches the interval used between the predictor and aversive stimuli (Davis
et al., 1989; Siegel, 1967). The neural circuits underlying fear potentiation of animal
startle have been well described (see Davis, 1996). Fear potentiation can be
observed when the lead stimuli and the startle stimuli are in different modalities,
which is not what would be expected viewing startle from an attentional framework. Affectively valenced stimuli appear to have unique effects on the startle
response.
In the first examination of affective processing and SEM in humans, Vrana et al.
(1988) presented participants with a series of six second pictorial slides depicting
affectively positive (e.g. smiling baby), negative (e.g. mutilated body) and neutral
(e.g. basket) images, and acoustic startle-eliciting stimuli were presented during the
slides at lead intervals of 500, 2500, and 4500 ms. These pictures, as well as those
used in the studies throughout this section, are from a standardized library of
images that have been calibrated for valence and arousal level (The International
Affective Picture System (IAPS); Center for the Study of Emotion and Attention,
1995; Lang et al., 1995). Thus it is important to note that the positive and negative
pictures were equally arousing and attention-demanding, based on these calibrations and on findings that heart rate orienting responses and free viewing time did
not differ between the affective picture categories. The results, re-presented in Fig.
6, revealed a significant linear trend at long (but not short) lead intervals over
picture valence, with the largest startle responses occurring during unpleasant
pictures and smallest during pleasant pictures.
This affective modification effect has been interpreted within a theoretical
framework focused on motivational priming (Lang et al., 1990). According to this
framework, reflexive responding is dependent on two factors: the classification of
the reflex as appetitive or defensive, and the affective valence of the individual’s
ongoing emotional state. Defensive reflexes such as startle will be enhanced if
elicited in the context of a negative emotional state, whereas appetitive reflexes will
be enhanced if elicited in a positive emotional state. According to Lang (1995), ‘‘In
20
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
Fig. 6. Affective modulation of startle. Blink magnitude to startle-eliciting stimuli presented while
participants view emotional pictures (adapted from Vrana et al., 1988).
an unpleasant stimulus foreground the brain is processing negative affective information; the relevant subcortical, aversive system circuitry is contacted; and a
defensive reflex is augmented... when appetitive information is the focus of processing, the startle reflex is expected to show relative diminution.’’ (p. 379)
Over the past decade the affective modulation of startle in humans has been
replicated in numerous studies and has gained much attention in the study of
human emotion, as evidenced by several recent reviews of the extensive emotional
modification of startle literature (Bradley et al., in press; Lang, 1995; Lang et al.,
1990, 1992, 1993). Affective modulation of startle has been shown to be quite
ubiquitous, occurring as early as 5 months of age (Balaban, 1995), and generalizing
to a variety of other lead stimulation conditions which affect the emotional state of
the participant, including pleasant and unpleasant film clips (Jensen and Frijda,
1994), pleasant and unpleasant odours (Ehrlichman et al., 1995; Miltner et al.,
1994), as well as fear-potentiation conditions in which the lead stimuli signal an
upcoming aversive event such as electric shock (e.g. Grillon et al., 1991; Hamm et
al., 1993). Moreover, in the fear-potentiation paradigm, startle facilitation has been
shown to increase as a function of anticipatory anxiety, with the largest facilitation
occurring just prior to delivery of the aversive stimulus (Grillon et al., 1993). In
addition, affective modulation of startle has also been demonstrated in situations in
which the lead stimulus serves as a cue for participants to engage in an affective
task such as rehearsing sentences with differing affective tone (Vrana and Lang,
1990) or imagining pleasant versus unpleasant scenes (Cook et al., 1991). Although
fear appears to have been the negative emotion most often studied, other negatively
valenced emotions such as sadness, anger, and disgust also produce potentiated
startle (Cook et al., 1991; Witvliet and Vrana, 1995; Vrana, 1994).
Consistent with the motivation priming hypothesis, Bonnet et al. (1995) found no
affective modulation of the spinal tendinous reflex, a reflex that is not inherently
defensive. The effect also seems to be particularly resistant to habituation (Bradley
et al., 1993a), and appears to be most robust in the context of highly arousing
stimuli (Cuthbert et al., 1996). In addition, Bradley et al. (1991, 1996) have reported
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
21
evidence suggesting that this emotion modification effect may be lateralized. In
these studies, participants were presented with startle-eliciting stimuli consisting of
left and right ear monaural noise bursts at lead intervals of 2000–6000 ms while
viewing positive, negative, and neutral images. The results revealed that startle
stimuli presented to the left ear (activating predominantly the right hemisphere)
produced a significant pattern of modulation by the affective nature of the pictures
whereas reflexes elicited by startle stimuli presented to the right ear (activating
predominantly the left hemisphere) showed no significant emotional modulation.
This finding is consistent with current theories suggesting that the right hemisphere
plays a dominant role in the processing of emotional information (e.g. Bowers et
al., 1993). However, Hawk and Cook (1997) found no laterality effect with tactile
startle stimuli.
Finally, as was the case with the short lead interval effects, in which startle
inhibition deficits in various clinical populations appear to validate the protection
of processing/sensorimotor gating hypothesis, there are also clinical data which
appear to validate the motivational priming interpretation of the emotional modulation effects. As will be discussed in more detail later, Patrick et al. (1993) found
that criminal psychopaths failed to show the typical linear relationship between lead
stimulus valence and SEM, with the lack of emotional modulation observed
specifically in prisoners who, at interview, showed emotional indifference and lack
of remorse. Also, individuals with specific phobias have been reported to exhibit a
larger startle facilitation effect while viewing pictures of their own phobic objects
than do non-phobic individuals viewing these or other unpleasant pictures (Hamm
et al., 1997).
The studies reviewed in this section clearly illustrate that the emotional context in
which startle is elicited can modulate the amplitude of the startle reflex. Although
this phenomenon has been replicated many times across a variety of experimental
protocols, several important questions remain unanswered. For example, what are
the roles of arousal and attention in affective modulation of startle? These
questions are particularly germane in light of recent reports that affective modulation of the startle amplitude occurs only with affective stimuli that are highly
arousing. Pictures which were rated lower in arousal, although they differed in
affective valence as much as did the high arousal pictures, did not produce affective
modulation of startle (Cuthbert et al., 1996). The research of Cook and his
colleagues, reported in detail below, which indicates greater emotional modulation
of startle among high fearfulness than low fearfulness individuals, is also relevant to
this issue. For instance, Cook et al. (1991) reported that low fear participants, for
whom arousal during negatively valenced imagery would presumably be lower than
for high fear subjects, did not show significant emotional modulation of startle. On
the other hand, Witvliet and Vrana (1995) reported affective modulation of startle
using sentences having emotional content that were low in arousal as well as high
in arousal. Paradigms that allow the simultaneous measurement of both affective
and arousal/attentional processes are relevant to these issues and are beginning to
be published, as discussed in the following section.
22
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
4.3. Attentional and affecti6e effects compared
The studies reviewed above examining the effects of attentional processing on
SEM provide support for the attentional resource framework in explaining/predicting long lead interval SEM effects. Likewise, the studies reviewed above examining
the effects of emotional processing on SEM provide support for the motivational
priming framework in explaining/predicting long lead interval SEM effects in
situations in which the lead stimuli possess a strong affective valence. Thus, the
attentional resource framework suggests that the size of the startle reflex will be
determined by the amount of attention allocated to the modality of the startle
stimulus, and the motivational priming framework suggests that the size of the
startle reflex will be determined by the affective valence of the lead stimulus. What
is not well understood at this point is the relationship between these frameworks.
Are both frameworks necessary or is it possible that one of these frameworks may
be capable of accounting for both sets of findings? To date, five studies have
addressed this issue.
In the first study, Bradley et al. (1990) noted that the basic attention-related effect
of the Anthony and Graham (1983, 1985) studies of smaller acoustic startle
responses during interesting visual lead stimuli could be easily accounted for by the
motivational priming framework by simply making the assumption that the interesting lead stimuli had a high positive affective valence. In addition, they noted that
the basic emotional modulation findings with acoustic startle reflexes could be
accounted for within the attentional resource framework by making the assumption
that the negative lead stimuli resulted in attention being directed away from the
visual modality. Based on this latter assumption, they designed an experiment in
which the attentional hypothesis and the motivational priming hypothesis made
competing predictions. This experiment involved lead stimuli consisting of positive,
negative, and neutral pictures and both acoustic and visual startle-eliciting stimuli.
Based on the assumption that lead stimuli consisting of unpleasant pictures would
result in attention being directed away from the visual modality, the attentional
resource hypothesis predicts that for visual startle stimuli, unpleasant pictures
should produce smaller reflexes than positive pictures. In contrast, the motivational
priming hypothesis predicts the opposite. The results supported the motivational
priming hypothesis; over lead intervals ranging from 2500 to 5500 ms, startle
eyeblinks were largest during the unpleasant pictures and smallest during the
pleasant pictures (showing the same pattern as shown in Fig. 6), regardless of the
modality of the startle-eliciting stimulus.
In a second study, Bradley et al. (1993a) tested the hypothesis that the startle
reflex may be sensitive to both attentional and emotional processes, but that the
time course of these processes may differ. Participants were again presented with
positive, negative, and neutral pictures and auditory startle-eliciting stimuli, but
instead of presenting the startle stimulus at lead intervals ranging from 2500 to 5500
ms, the startle stimulus was presented at lead intervals of 300, 800, 1300, and 3800
ms following picture onset. The results revealed a significant effect of attention at
the 300 ms lead interval; i.e., startle responses were smaller during both pleasant
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
23
and unpleasant pictures than during the neutral pictures. Bradley et al. (1993a)
interpreted this effect as due to attentional processes, not emotional processes,
because positive and negative pictures, which were both more interesting than the
neutral pictures, had identical effects. Emotion effects were observed at all three of
the long lead intervals, with startle responses smaller during pleasant pictures than
unpleasant pictures. Bradley et al. (1993a) concluded that these results provide
support for the motivational priming hypothesis in accounting for long lead interval
SEM effects, and suggested that attentional influences account for short lead
interval effects.
In a third study, Robinson and Vrana (1995) noted a critical difference in the
way attention was defined in studies with affectively loaded stimuli (e.g., Bradley et
al., 1993a) and those with neutral stimuli (e.g., Filion et al., 1993). In the Bradley
et al. (1993a) study, attention was determined by the affective arousal properties of
the pictures themselves, with both pleasant and unpleasant pictures assumed to be
more attention-engaging than the neutral pictures. In the Filion et al. (1993) study,
attention was manipulated by instructing participants to estimate the duration of
one class of affectively neutral lead stimulus and to ignore another class. Robinson
and Vrana noted that one way to manipulate attention and emotion independent of
the qualities of the lead stimulus is to use a neutral lead stimulus as a signal for an
emotional task. Participants, therefore, were presented three tones varying in pitch
with the high and low tones serving as signals to generate pleasant and unpleasant
images, and auditory startle probes presented at short and long lead intervals
following tone onsets. Tones signalling emotional imagery produced greater inhibition of startle at the short 120 ms lead interval than did tones not signalling the
imagery task, consistent with a short lead interval attention effect, but the two
tones signalling pleasant and unpleasant imagery did not produce differential short
lead interval inhibition. However, the tone that cued negatively valenced emotional
imagery did produce larger startle magnitude at long intervals of 1400 ms than the
tone signalling positively valenced emotion. Thus the results corroborate earlier
findings that attention can influence startle at short intervals and affective modulation occurs later.
In a fourth study, Vanman et al. (1996, experiment 2) devised an ‘affect-directed
attention’ paradigm to independently manipulate attentional and emotional modulation of startle with affectively loaded pictures. Participants were instructed to
either count the longer than usual occurrences of all the pleasant pictures and
ignore all unpleasant pictures or the opposite, with the attended valence counterbalanced across participants. Thus the emotion quality of the picture became an
attentional cue to perform or not perform the duration judgment task. Therefore,
across participants, startle was recorded during attended and unattended pictures
with both positive and negative affective tone. Auditory startle probes were
presented at lead intervals of 250, 750, and 4450 ms following picture onsets, as well
as 950 ms following picture offsets. Affective modulation, with larger startle blinks
following the unpleasant pictures than following the pleasant pictures, was found at
the both short lead intervals (250 and 750 ms) and the long lead interval (4450 ms).
In contrast, the only evidence of attentional modulation occurred at the lead
24
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
interval 950 ms following picture offset, where the offset of the attended pictures
produced greater startle inhibition than the offset of the ignored pictures. This is
the first study to demonstrate affective modulation of the human eyeblink at short
lead intervals, although a similar effect has been reported with rats in a fear
conditioning paradigm (Davis et al., 1989). Vanman et al. (1996) concluded that
both affective and attentional modulation of startle can occur at both short and
long lead intervals depending upon task requirements and stimulus parameters.
Taking a very different approach, Haerich (1994) also examined the relative
contributions of attentional and affective processes to SEM. In this experiment,
participants were presented with a series of trials in which a visual warning stimulus
was followed at a lead interval of 6 s by either an eyeblink-eliciting air puff or an
eyeblink-eliciting noise burst, with presentation order unpredictable from the
participant’s point of view. For half of the trials, participants were instructed to
focus their attention on and judge the length of one of these two startle-eliciting
stimuli and ignore the other, with the instruction reversed on the remaining trials.
In this way, the influences of attentional focus on startle responding could be
assessed. In addition, half of the participants in this experiment heard descriptions
of the air puff stimulus that suggested a negative emotional context, (i.e. ‘the air
puff will be directed toward your eye’), whereas the remaining participants received
neutral information (i.e. ‘the air puff will be directed near your ear’). This
manipulation allowed an assessment of the effect of emotional context on SEM,
although in fact the location of the air puff was identical in both groups. The
results revealed that in the neutral-context group, startle responding varied directly
with attentional focus, with larger responses elicited when the startle-eliciting
stimulus was presented in the attended modality. In contrast, participants in the
negative context group showed larger startle responses to the airpuff eliciting
stimulus, regardless of the focus of attention.
The five studies reviewed above demonstrate that both affective and attentional
modulation of startle can be observed within the same participants within the same
paradigm. Paradigms such as these may provide a window on the interactions and
temporal relationships between affective and attentional processes, potentially
disentangling their mutual effects. For example, this line of research may be able to
determine whether certain types of stimuli elicit emotions pre-attentively (i.e.
Zajonc, 1980), or whether the elicitation of emotions requires prior attention and
cognitive evaluation of the stimulus (i.e. Lazarus, 1984).
5. Application of startle modification in clinical research
SEM paradigms are being increasingly applied in clinical research. Although
these applications have involved a wide variety of clinical populations, we will focus
here on the use of the short lead interval inhibition paradigm in the study of
schizophrenia spectrum disorders, and on the use of the long lead interval affective
modulation paradigm in the study of affective disorders and psychopathy.
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
25
5.1. Short lead inter6al startle inhibition
As described earlier, short lead interval startle inhibition is widely viewed as a
measure of the protection of early stimulus processing. For this reason, startle
inhibition measures have been of particular interest in the study of schizophrenia
because deficits in early stimulus processing are believed to represent core characteristics of the disorder, possibly related to underlying vulnerability factors (Cadenhead and Braff, in press). To date, there are five published reports of startle
inhibition in schizophrenia (Bolino et al., 1994; Braff et al., 1978, 1992; Dawson et
al., 1993; Grillon et al., 1992).
In the earliest of these studies (Braff et al., 1978), hospitalized inpatients with
schizophrenia were tested in a passive attention paradigm (participants were given
no explicit task to perform) involving a continuous pure-tone lead stimulus and a
white noise startle-eliciting stimulus presented at short lead intervals ranging from
30 to 500 ms. The results revealed that the schizophrenia patients exhibited
significantly less startle inhibition than controls at a 60 ms lead interval. Braff et al.
(1992), also studying inpatients, employed a similar paradigm using a discrete white
noise burst as the lead stimulus and found significantly reduced startle inhibition in
the schizophrenia inpatients at lead intervals of 30, 60, and 120 ms. Grillon et al.
(1992) also observed reduced startle inhibition in schizophrenia patients at a 120 ms
lead interval using a discrete white noise lead stimulus presented at intensities
ranging from 75 to 90 dB(A). In addition, in order to determine whether this
schizophrenia-linked deficit is modality specific, Braff et al. (1992) included a
condition in which the startle-eliciting stimulus was tactile rather than acoustic and
found that the schizophrenia patients exhibited deficient startle inhibition in the
tactile condition as well. These results have been interpreted as indicating that
patients with schizophrenia have poor automatic, pre-attentive sensorimotor gating
that can lead to sensory overload, thought disorder, and perhaps the cognitive
fragmentation associated with severe psychotic symptoms (e.g. Braff et al., 1995).
Two more recent studies, however, have revealed a more complex pattern of results.
Bolino et al. (1994) examined SEM in hospitalized schizophrenia inpatients using
an electrocutaneous startle-eliciting stimulus (shock to the forehead activating the
trigeminal nerve) and discrete electrocutaneous lead stimuli at short lead intervals
of 30, 60, and 120 ms. With electrocutaneous stimuli, two distinct EMG components are generally observed: R1, generally occurring in a latency window less than
25 ms, and R2, occurring later and typically showing an inhibitory effect of lead
stimulation that is most closely associated with the eyeblink measured to acoustic
and visual eliciting stimuli. In contrast to the previously reviewed findings (e.g.
Braff et al., 1978), Bolino et al. found no differences between schizophrenia and
control participants in terms of the inhibition of the R2 component of the startle
response. However, there was a difference between the groups in the temporal
pattern of Rl facilitation. Both groups showed facilitation of R1 magnitude at the
60 ms lead interval, but only the control group sustained that facilitatory effect at
the 120 ms lead interval. The authors conclude that the facilitory neural pathways
are abnormal in schizophrenia patients.
26
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
Dawson et al. (1993) tested patients with schizophrenia and matched normal
controls in the active attention tone-length judgment paradigm described previously (e.g. Filion et al., 1993). The schizophrenia participants were outpatients
and were or had been participants in a longitudinal study of the early phases of
schizophrenia (Nuechterlein et al., 1992). These patients were in the relatively
early phases of schizophrenia without a long history of medication and institutionalization and were relatively asymptomatic at the time of testing, as assessed
by independent ratings on an expanded version of the Brief Psychiatric Rating
Scale (Overall and Gorham, 1962; Lukoff et al., 1986) Normal control participants also were drawn from participants in the longitudinal research project and
were matched to the outpatients on age, sex, race, and educational level.
In this active attention paradigm, participants were presented an intermixed
series of tones of two different pitches and were instructed to count the number
of longer than usual occurrences of one particular tone and to simply ignore the
other tone. In addition to the tone stimuli, a startle-eliciting noise burst was
presented at short lead intervals of 60, 120, and 240 ms following the to-be-attended tones and the to-be-ignored tones. Recall that, using this task, college
student participants have been reported to exhibit significantly greater inhibition
following the attended tone than the ignored tone at the 120 ms lead interval,
and significant but non-differential inhibition during the attended and ignored
tones at the 60 ms lead interval (e.g. Filion et al., 1993). This pattern of results
led Dawson et al. (1993) to suggest that startle inhibition at the 60 ms lead
interval reflects predominantly automatic sensorimotor gating, whereas inhibition
at the 120 ms lead interval reflects the additional allocation of controlled attentional resources. The results of the Dawson et al. (1993) study revealed that the
schizophrenia patients failed to show attentional modulation of startle inhibition
at 120 ms but exhibited normal inhibition at 60 ms, leading to their conclusion
that ‘‘relatively asymptomatic outpatients are not measurably deficient in the
early pre-attentive stimulus detection and evaluation process—rather, they are
deficient in the allocation of controlled resources at 120 ms to evaluate the
to-be-attended prepulse…’’ (p. 639).
Although these results and conclusions, as well as those of Bolino et al.,
appear inconsistent with those of the earlier studies, there are several key
methodological differences that may account for the apparent discrepancies, including the characteristics of the research participants. The deficits in automatic
startle inhibition observed in the Braff et al. (1978, 1992) studies were observed
in chronic hospitalized inpatients, whereas deficits in automatic startle inhibition
were not observed by Dawson et al. (1993) in their sample of remitted outpatients.
One strategy for increasing our understanding of the information processing
deficits associated with schizophrenia that reduces the complication of chronicity
and hospitalization is to examine SEM in ‘at-risk’ populations (individuals who
are considered to be vulnerable to psychosis but who are not psychotic). This
strategy was employed by Swerdlow et al. (1995a), who examined startle inhibition in ‘normal’ participants recruited from the community whose profile on the
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
27
Minnesota Multiphasic Personality Inventory (MMPI) is considered ‘psychosisprone’ based on previously published theoretically and empirically derived criteria
(Butler et al., 1993). Using a passive attention startle modification paradigm with
a lead interval of 60 ms, Swerdlow et al. reported that the MMPI-defined
‘psychosis prone’ group exhibited significantly less startle inhibition compared to
controls. A similar strategy was also employed by Cadenhead et al. (1993), who
examined startle inhibition in individuals meeting the DSM-III-R criteria for
schizotypal personality disorder. These patients manifest symptoms in the
schizophrenia spectrum, but to a less severe degree than those diagnosed with
schizophrenia. Cadenhead et al. (1993) also employed a passive attention
paradigm with acoustic lead and startle-eliciting stimuli and found that individuals with schizotypal personality disorder exhibited significantly less startle inhibition than controls at lead intervals of 30, 60 and 120 ms. In contrast to the
findings of Braff et al. (1992) with schizophrenia patients, however, no differences
between groups were observed in a condition involving a tactile startle-eliciting
stimulus.
A third putatively vulnerable ‘psychosis-prone’ population includes individuals
who experience extreme Perceptual Aberrations and/or Magical Ideation (Chapman and Chapman, 1987; Chapman et al., 1994). Individuals who score high on
true-false questionnaires designed to evaluate these characteristics are called ‘PerMags’, and these individuals have been shown to exceed normal scoring control
participants on rate of diagnosis of psychosis at a 10 year follow up (Chapman
et al., 1994). There have been six studies to date that have examined startle
inhibition in Per-Mag individuals (see Dawson et al., 1995 for a review).
Employing a passive attention paradigm, Simons and Giardina (1992) found
that Per-Mag college students exhibited less startle inhibition than normal controls at a lead interval of 120 ms, but exhibited comparable inhibition at a lead
interval of 60 ms. Perlstein et al. (1989) found no Per-Mag deficit at a lead
interval of 120 ms but did observe reduced startle inhibition at a 500 ms lead
interval. In contrast to these positive results, no differences in startle inhibition
between Per-Mags and control participants were observed by Cadenhead and
Braff (1992) at 30, 60 or 120 ms lead intervals, nor by Blumenthal and Creps
(1994) at lead intervals of 60 or 120 ms, nor by Lipp et al. (1994) across a wide
variety of lead intervals, including 120 ms. Finally, Schell et al. (1995) tested
Per-Mag college students in the active attention tone-length judging paradigm
described previously with short lead intervals of 60, 120 and 240 ms. As in the
previous studies, the normal controls exhibited startle inhibition at each of the
short lead intervals and enhanced startle inhibition at the 120 ms lead interval
following the attended compared to the ignored lead stimulus. In contrast, the
Per-Mag participants failed to show this attentional modulation of startle inhibition at 120 ms although they did show enhanced startle inhibition during the
attended lead stimulus at 240 ms, suggesting that attentional processing may
follow a slower time course in these individuals. The Per-Mags also showed no
evidence of deficits in automatic startle inhibition (i.e. startle inhibition was
normal at the 60 ms lead interval).
28
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
5.2. Clinical applications of short lead inter6al SEM: conclusions
Together, the short lead interval SEM studies described above present a complex
picture regarding the nature of information processing dysfunctions in schizophrenia and at-risk populations. Although some of the apparent discrepancies in the
studies of schizophrenia appear to be at least partially due to the characteristics of
the participants (i.e. chronic/acute, psychotic/remitted), there are also apparently
discrepant findings in the at-risk populations, even within similarly defined participants. For example, five of the studies reviewed above examined SEM in Per-Mag
college students utilizing a passive attention paradigm. Three of these studies
observed no Per-Mag deficit in startle inhibition across lead intervals of 30, 60, and
120 ms (Blumenthal and Creps, 1994; Cadenhead and Braff, 1992; Lipp et al.,
1994), whereas the other two did observe deficits, one at a lead interval of 120 ms
(Simons and Giardina, 1992), and one at a lead interval of 500 ms (Perlstein et al.,
1989). Moreover, Schell et al. (1995), utilizing an active attention paradigm,
observed no Per-Mag deficit in automatic startle inhibition, and observed that
attentional modulation of startle inhibition followed a different time course for the
Per-Mag group than the control group, with controls exhibiting the attentional
effect at a 120 ms lead interval and the Per-Mag group showing the effect at a lead
interval of 240 ms.
Thus it is clear that differences in characteristics of the research participants are
not sufficient to account for all of the complexities in this literature, and that
characteristics of the lead and startle-eliciting stimuli as well as of the tasks and lead
intervals are also critical factors. For example, Braff and colleagues have reliably
observed a schizophrenia-related deficit using acoustic and tactile startle-eliciting
stimuli, whereas Bolino et al. observed no such deficit using an electrically elicited
R2 response. In addition, in terms of the characteristics of the SEM paradigm used,
deficits in automatic startle inhibition were observed using a passive attention
paradigm in the studies by Braff and colleagues, whereas only deficits in the
attentional modulation of startle inhibition were observed by Dawson et al. (1993).
One possible explanation for this latter discrepancy suggested by Dawson et al.
(1997) is that ‘‘Requiring active attention may alter the underlying processes such
that automatic deficits are no longer detectable; instead only controlled processing
impairments are apparent’’. Thus it may be the case that schizophrenia involves
deficits in both automatic sensorimotor gating and controlled processing, but that
the introduction of a task increases the overall activation level of schizophrenic
patients and subsequently increases the level of ‘automatic’ startle inhibition.
Moreover, it may be possible that the nature of the SEM deficit in schizophrenia
changes over the developmental course of the disorder. An important direction for
future research into SEM and schizophrenia will be to clarify these factors.
5.3. Long lead inter6al startle modification
As described earlier, SEM at long lead intervals has been shown to be sensitive
to emotional processes; negative or unpleasant lead stimuli produce startle facilita-
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
29
tion, whereas positive/pleasant lead stimuli produce startle inhibition. (See review
by Cook, in press). For this reason, affective SEM paradigms are becoming
increasingly popular in the study of affective disorders and psychopathy. Current
research in these areas will be discussed below.
One application of the affective modulation paradigm has been focused on the
study of anxiety states such as fearfulness, phobias, and stress disorders. In the first
of such studies, Cook et al. (1991) examined SEM in college students classified into
high-fear and low-fear groups based on their responses to a version of the Fear
Survey Schedule. Participants performed an imagery task in which they were cued
on different trials to imagine joyful, relaxing, fearful, angry, sad, and neutral
scenes/activities, and a startle-eliciting noise burst was presented at lead intervals
ranging from 8 to 13 s into the imagery period. The results revealed that startle
responses were significantly larger during negative than positive imagery, and that
this effect was greater in high-fear than low-fear participants. Startle potentiation
was not specific to fear; rather it occurred during all of the negative affect imagery
(sadness, anger, and fear). In a second study, Cook et al. (1992) examined startle
responding in participants viewing a series of intermixed neutral (plants, household
objects, etc.) and aversive (wounds, mutilated bodies, etc.) pictures. A warning
stimulus informed participants as to the content of each picture presented, and a
startle-eliciting noise burst was presented while participants were viewing the
pictures. The results revealed that high fear participants exhibited significantly
larger startle responses while viewing the aversive pictures than did the low fear
participants. Together, these results suggest that startle facilitation can discriminate
high and low fear participants on the basis of their differential responding to
positive and negative stimuli, but may be limited in discriminating specific emotions
within the negative category such as fear, anger, and sadness.
A related area of research in which the affective modulation of SEM technique
has been utilized is in the study of phobias. Hamm et al. (1997) have reported that
individuals show a significantly larger startle facilitation effect when viewing
pictures of their specific phobic object than when viewing other unpleasant pictures
and greater facilitation while viewing the phobic objects than non-phobic individuals viewing the same or other unpleasant pictures. Moreover, this phobia-related
facilitation effect is so robust that it has been proposed as an instrument for
evaluating the effectiveness of specific phobia therapies (de Jong et al., 1993). de
Jong et al. (1991) measured startle responses while spider phobics were viewing live
spiders or food items before and after one exposure and modelling treatment
session. Initially, startle responses were greater during viewing of the spiders than
the food items, as would be expected. After the treatment session, although
responses were still larger during spider viewing, the degree of emotional modulation of startle had declined significantly. Similar results were obtained in two case
studies by Vrana et al. (1992), who found that a series of systematic desensitization
treatments eliminated the initial difference in startle response during phobic and
neutral imagery. A somewhat different pattern of results was obtained by de Jong
et al. (1993), who reported that after one treatment, startle responses during
viewing of phobic, pleasant, and neutral items all declined. Emotional modulation
30
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
of startle was not significant before or immediately after treatment, but was at a
one-week follow-up. The initial absence of emotional modulation of startle was
attributed to a generalized fear-induced potentiation, since participants were aware
that phobic objects (live spiders) would be viewed.
Although Hamm et al. (1997) found that persons with specific phobias show
larger startle facilitation when viewing their phobic objects that other unpleasant
pictures, results from an investigation that included participants with simple
phobias, social phobia, post-traumatic stress disorder (PTSD), and panic disorder
suggest that not all anxiety states exhibit this same pattern of results (Lang, 1995).
In this study, participants read a sentence that involved either routine and affectively neutral events, anxiety evoking social situations (an embarrassing act or
speech performance), situations involving threat or danger (an auto accident or a
night time intruder), or situations directly relevant to their specific problem.
Participants were instructed to imagine themselves in the situation described by the
sentence. Startle-eliciting stimuli were then presented at long lead intervals during
the imagery task. The results revealed that all participants showed startle potentiation in the threat-danger condition compared to the neutral condition. However,
whereas the participants with simple and social phobias also showed startle
potentiation during imagery involving their specific phobic object situation, participants in the panic and PTSD groups did not exhibit startle potentiation during
imagery involving their specific traumatic events. These data suggest that the startle
facilitation measure may be valuable in understanding the dynamics of anxiety
disorders.
A third area of anxiety research in which the SEM paradigm has been applied is
in the area of fear-potentiation in PTSD. In order to test the hypothesis that
individuals with PTSD might show an exaggerated fear-potentiated startle response,
Morgan et al. (1995) conducted an experiment in which startle-eliciting stimuli were
presented to PTSD and control participants first during a simple habituation series
and then during alternating shock-threat and no-threat anticipatory intervals.
During the latter intervals participants were presented with two coloured lights, one
of which reliably predicted the presentation of a shock to the forearm, the other of
which signalled a no shock-threat period. Each light was presented for 50–60 s, and
startle-eliciting stimuli were presented at lead intervals ranging from 5 to 45 s.
Participants were instructed that when a shock was presented it would always be
presented during the last 10 s of a trial. The results of this experiment revealed that
both PTSD and control participants exhibited startle facilitation during the shockthreat intervals, with the amount of facilitation increasing as the likelihood of shock
increased. Contrary to prediction, the individuals with PTSD did not show an
exaggerated facilitation of startle during the shock-threat condition, but rather
exhibited significantly larger startle responses than controls across all three conditions (habituation, threat, and no-threat conditions). One possibility suggested by
Morgan et al. (1995) is that the increased baseline responding may have been due
to an increase in generalized anxiety in the PTSD individuals brought on by the
stress of participating in an experiment involving an aversive stimulus.
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
31
Another promising avenue for SEM research and affective processes is the area
of psychopathy. Cleckley (1976) provided the classic description of the psychopath
and emphasized the importance of general poverty of affect in this disorder (e.g.
lack of remorse, absence of nervousness, inability to love, and failure to establish
close relationships). Therefore, one would predict that psychopathy would be
characterized by poor affective modulation of SEM. Patrick et al. (1993) (see also
Patrick and Lang, in press) tested this prediction by examining SEM in prison
inmates while they viewed pictures of pleasant, neutral, and unpleasant pictures.
The results for inmates without a psychiatric diagnosis were comparable to those
typically observed with this paradigm; startle responses were largest during the
unpleasant pictures and smallest during the pleasant pictures. In contrast, the
criminal psychopaths showed no significant startle facilitation during the viewing of
unpleasant pictures. Moreover, this lack of emotional modulation was observed
specifically in prisoners who, at interview, showed emotional indifference and lack
of remorse as opposed to only a history of criminal behaviour. Thus the lack of
affective modulation of startle responding confirmed the clinical assessment of
reduced emotional reactivity. Patrick (1994) has reported a replication of these
results.
In contrast to this finding, however, a recent study of schizophrenia patients
found a lack of agreement between clinical/behavioral assessments of emotional
reactivity and startle responding (Schlenker et al., 1995). In this experiment,
schizophrenia inpatients viewed a series of pleasant, unpleasant, and neutral
pictures from the IAPS, and startle-eliciting stimuli were presented during the
pictures at lead intervals of 3.1 and 5.1 s. Using video tapes of clinical interviews,
the patients were rated on a dimension termed ‘diminished affective expression’
based on items such as paucity of expressive gestures, unchanging facial expression,
and lack of vocal expressions. Patients scoring high on this dimension were then
compared to low-scorers in terms of their affective modulation of startle. The
results revealed that patients rated as having a highly diminished affective expression did not differ from controls; both groups showed the typical linear response
pattern of startle facilitation during unpleasant slides and startle inhibition during
pleasant slides. Surprisingly, patients for whom affective expression was not diminished failed to show this pattern, instead exhibiting equally reduced startle responding during both pleasant and unpleasant pictures. Also surprising was the finding
that affective modulation of startle in patients who had high scores on a self-assessed anhedonia survey did not differ from that seen in low anhedonia patients,
and that neither of these subgroups displayed the typical ‘unpleasant\ neutral\
pleasant’ pattern of startle responding. Thus these results indicate a dissociation
between behavioral assessments of emotional reactivity, subjective reports of emotional experience, and affective modulation of startle in schizophrenia. Thus SEM
may provide unique information about affective processes that is not available from
verbal and behavioral measures. Dissociation of expressive, experiential, and psychophysiological (skin conductance) measures of emotion also has been reported
with schizophrenia patients by Kring and Neale (1996), further confirming the need
for measuring multiple components of emotion.
32
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
5.4. Clinical applications of long lead inter6al SEM: conclusions
Together, the studies reviewed above suggest that SEM at long lead intervals is
sensitive to individual differences in affective states and traits in both the normal
and pathological range. High fear college students exhibit significant affective
modulation of startle, with larger responses during negative affect than positive
affect, to a significantly greater degree than low fear college students. Moreover,
individuals with simple and social phobias, PTSD, and panic disorder all exhibit
startle potentiation while viewing pictures or imagining scenes of generic threatdanger conditions, but only the phobic groups exhibit startle potentiation while
viewing pictures or imagining scenes related to their specific stressful object event.
Emotionally detached criminal psychopaths fail to potentiate the startle eyeblink
reflex during exposure to unpleasant pictures. There is growing evidence that
affective modulation of startle may provide information that is not available from
verbal reports and behavioral measures (e.g. startle modulation during emotional
pictures occurs in schizophrenia patients with diminished affective expression, but
not in patients without this behaviourally apparent deficit). All in all, the results
suggest that affective modulation of startle may be a powerful tool for detecting,
quantifying, and eventually understanding affective individual differences, particularly those characterized by anxiety and negative affect.
Important directions for future research include determining to what extent
affective modulation of SEM can be informative about (1) low arousal negative
emotions (e.g. flat affect, anhedonia, and depression), (2) individual differences in
positive emotions in addition to negative emotions, (3) identification of individuals
at-risk for emotional disorders, (4) distinguishing different emotional states (e.g.
fear versus anger) as well as different valences (pleasant versus unpleasant), (5) the
development of emotions across the lifespan, and (6) the modifiability of emotions
by learning, drugs, etc. (for work in this area, see Stritzke et al. (1995) regarding the
effects of alcohol on startle and its affective modulation and Patrick et al. (1996)
regarding the effects of diazopam). These directions of research can be profitably
pursued by a combined study of clinical and non-clinical human populations, as
well as of nonhuman animals.
6. Summary, conclusions, and unanswered questions
The present paper has reviewed psychological interpretations of the significance
of human startle eyeblink amplitude modification, with separate discussions of the
modification that occurs at short and long lead intervals. Table 1 summarizes the
principal psychological processes that have been hypothesized to mediate or
influence SEM at short and long lead intervals.
Startle eyeblink modification at short lead intervals has been hypothesized to be
associated with pre-attentive protection of processing (Graham, 1975, p. 246), with
the pre-attentive ‘call’ for attentional processes (Filion et al., 1994, p. 76), with
sensorimotor gating (Braff and Geyer, 1990), with the allocation of controlled
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
33
attentional processes (Dawson et al., 1993 p. 639), and most recently with affective
modulation (Vanman et al., 1996). It is important to note that these hypotheses
clearly are not mutually exclusive. Indeed, inhibition of startle amplitude at short
lead intervals is probably multiply determined and numerous psychological processes may be influential depending upon the nature of the lead stimulus, the length
of the short lead interval, the speedie task and state of the participants, and the
characteristics of the participants.
The long lead interval effects also have been hypothesized to be modulated by a
number of different processes, as summarized in Table 1: ‘classical activation’
(Graham, 1975, p. 243; see also Ison and Hammond, 1971, p. 449); non-specific
enhancement of sensory processing (Bohlin et al., 1981, p. 609), modality-specific
enhancement or inhibition of sensory processing (Bohlin et al., 1981, p. 609;
Graham, 1979, p. 160, Putnam, 1990, p. 119); and match/mismatch of affect/motivation (Lang et al., 1990). As with short lead interval SEM, multiple psychological
processes probably co-exist during long lead intervals, with different processes
being dominant depending upon characteristics of the stimuli, the task, and the
participants.
As can be seen, the psychological constructs hypothesized to mediate SEM
include both cognitive and affective/motivational processes at both short and long
lead intervals. The proposed cognitive processes consist of both pre-attentive
automatic processes and controlled attentional processes. As shown in Fig. 7, we
have previously suggested that the hypothesized effects of automatic and controlled
processes can be separately evaluated at both short and long lead intervals within
the differential tone length judgment task (Dawson et al., 1997). At short lead
intervals, we have noted that instructionally induced attentional processes do not
affect startle inhibition at 60 ms, but do have significant effects at 120 ms. Based on
these observations, we have proposed that within this paradigm startle inhibition at
the 60 ms lead interval represents automatic, pre-attentive processes, whereas startle
inhibition at 120 ms represents a combination of automatic and controlled attentional processes. According to this interpretation, the difference in startle inhibition
observed following the to-be-attended and to-be-ignored lead stimuli at 120 ms may
be used to index early controlled attentional processes, as may be the enhancement
of inhibition across the 60 to 120 ms lead intervals following the to-be-attended
Table 1
Psychological processes hypothesized to mediate SEM
SEM at Short Lead Intervals
SEM at Long Lead Intervals
Preattentive protection of processing
Preattentive ‘call’ for allocation of controlled
attention
Sensorimotor gating
Classical activation
Non-specific sensory enhancement
Allocation of controlled attention
Affective modulation
Modality-specific sensory
enhancement/inhibition
Affective modulation
34
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
Fig. 7. Hypothesized processes involved in startle eyeblink modification (adapted from Dawson et al.,
1997).
lead stimuli. At long lead intervals (e.g. 2000 ms), facilitation occurs following both
the to-be-attended and to-be-ignored lead stimuli, with greater facilitation following
the to-be-attended lead stimulus. These findings have led us to propose both a
non-specific facilitation effect, as indexed by the facilitation during the to-be-ignored lead stimulus, and controlled sustained attentional processes indexed by the
difference between the to-be-attended and to-be-ignored lead stimuli at the long
interval.
There is much that remains to be learned about the precise cognitive processes
affecting startle at short lead intervals, even within the specific differential tone
length judgment task. For example, although we know that attention directed
toward the lead stimulus can augment startle inhibition at the 120 ms lead interval
in this paradigm, what is the precise cognitive process responsible for this effect? On
different occasions, we have proposed that this effect reflects the pre-attentively
initiated ‘call’ for attention (Filion et al., 1994), or the answer to that call with
allocation of controlled attention (Dawson et al., 1993). These alternative hypotheses are currently being tested in our laboratory with two different strategies. First,
participants are engaged in an irrelevant task during the SEM paradigm, so that the
ability to answer the call with allocation of attentional resources is impaired but the
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
35
pre-attentive call is left intact (Seljos et al., 1995). Second, the controlled attentional
orienting response to the prepulse is habituated by repetitive presentation prior to
testing for its inhibitory power (Wynn et al., 1996).
Basic issues also remain concerning the cognitive processes affecting startle at
long lead intervals. For example, does the small but significant facilitation following
the to-be-ignored lead stimulus in the differential tone judgment task reflect
‘classical activation’ or non-specific sensory enhancement, and can these two
hypotheses be clearly differentiated? Is the facilitation observed at long lead
intervals in this paradigm following the to-be-attended lead stimulus associated
with modality-specific sensory enhancement? Recall that startle facilitation occurs at
long lead intervals if attention is directed toward the same modality as the startle
eliciting stimulus, whereas startle inhibition has been reported if attention is directed
to a modality different than that of the startle probe (e.g. Putnam, 1990). This
modality-specific effect suggests that the previously observed startle facilitation
following the to-be-attended acoustic lead stimulus (see Fig. 7) would become
startle inhibition if visual lead stimuli were used rather than auditory lead stimuli,
or if the startle reaction was elicited by visual stimuli rather than auditory stimuli.
However, contrary to this prediction, startle facilitation has been found in this
paradigm with both visual and acoustic to-be attended lead stimuli (Böhmelt et al.,
1996; Lipp et al., 1997). These results suggest that the long lead interval facilitation
effect in this paradigm is modality-nonspecific. Results such as these demonstrate
that our understanding of the basic psychological processes at long lead intervals
and their relationships to startle eyeblink modification, are in need of considerably
more research.
Similarly, much remains to be learned about affective modulation of startle with
the specific affectively loaded pictures typically used to study this phenomenon. For
example, can the affective valence effects be clearly separated from the effects of
arousal and attention on startle? How much of the affective modulation of startle
is due to the fact that highly affective pictures are also highly arousing and highly
interesting, compared to affectively neutral pictures? Is SEM differentially sensitive
to different emotions, or only to the bio-dimensional affective valence (positive
versus negative) associated with the emotion? That is, do all negatively valenced
emotions (fear, disgust, embarrassment, depression) potentiate startle? Do all
positively valenced emotions (happiness, pride, sexual arousal, interest) inhibit
startle? In other words, is the critical modulating process one of motivation and
valence, or does it have more to do with the varieties of emotions? Is the match
between the participant’s current motivational state and the affective quality of the
elicited reflex the underlying process that modulates startle? Is there independent
evidence for the emotional priming process hypothesized to cause affective modulation?
Although not emphasized in this review, a distinct advantage of studying human
SEM is that the same basic phenomena can also be investigated in non-humans.
Research with non-humans can more readily investigate the neurophysiological and
neurochemical basis of startle modification than can human research. In fact, the
study of startle modification in non-humans has already produced considerable
36
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
knowledge about the neurocircuitry of the primary acoustic startle reflex (e.g. Koch
et al., 1992; Lee et al., 1996), its modification at short lead intervals (e.g. see reviews
by Dawson et al., 1997; Swerdlow and Geyer, in press; Swerdlow et al., 1992), and
its affective modification at long lead intervals (e.g. Davis, 1996; Davis et al., in
press). Any complete explanatory model of startle modification must take into
account the neurophysiological and neurochemical basis of the phenomena, as well
as their psychological significance. Hackley (in press) has reviewed evidence for
models that incorporate cortical-subcortical pathways as well as conscious-unconscious processes. In this sense, a complete model of startle modification will be a
psychophysiological model. Psychophysiology lies at the interface of cognitive
science, neuroscience, and clinical science (e.g. Dawson, 1990), and the study of
startle modification has the potential to be a model psychophysiological system for
theoretical integration across disciplinary boundaries.
Acknowledgements
D. Filion was supported by grant MH55043, M. Dawson was supported by a
Research Scientist Development Award (1 KO2 MH1086), and both A. Schell and
M. Dawson were supported by grant MH46433, all from the National Institute of
Mental Health, during the preparation of this review. The authors wish to thank
Andreas Böhmelt and the reviewers for their helpful comments on an earlier draft
of the paper.
References
Acocella, C.M., Blumenthal, T.D., 1990. Directed attention influences the modification of startle reflex
probability. Psycholog. Rep. 66, 275– 285.
Anthony, B.J., 1990. Blink modulation in attention deficit disorder [abstract]. Psychophysiology 27
(Suppl.), S6.
Anthony, B.J., 1985. In the blink of an eye: implications of reflex modification for information
processing. In: Ackles, P.K., Jennings, J.R., Coles, P.G.H. (Eds.), Advances in Psychophysiology, 1.
JAI Press, Greenwich, pp. 167–218.
Anthony, B.J., Graham, F.K., 1983. Evidence for sensory-selective set in young infants. Science 220,
742–744.
Anthony, B.J., Graham, F.K., 1985. Blink reflex modification by selective attention: evidence for the
modulation of ‘automatic’ processing. Biol. Psychol. 21, 43 – 59.
Anthony, B.J., Putnam, L.E., 1985. Cardiac and blink reflex concomitant of attentional selectivity: a
comparison of adults and young children. Psychophysiology 22, 508 – 516.
Balaban, M.T., 1995. Affective influences on startle in five-month-old infants: Reactions to facial
expressions of emotion. Child Develop. 66, 28 – 36.
Balaban, M.T., Anthony, B.J., Graham, F.K., 1989. Pre-stimulation effects on blink and cardiac reflexes
of 15-month human infants. Develop. Psychobiol. 22, 115 – 127.
Balaban, M.T., Losito, B., Simons, R.F., Graham, F.K., 1986. Offline latency and amplitude scoring of
the human reflex eyeblink with FORTRAN IV [abstract]. Psychophysiology 23, 612.
Berg, K.W., Balaban, M.T., in press. Startle elicitation: stimulus parameters, recording techniques, and
quantification. In: Dawson, M.E., Schell, A.M., Böhmelt, A.H. (Eds.), Startle Modification: Implica-
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
37
tions for Neuroscience, Cognitive Science, and Clinical Science. Cambridge University Press,
Cambridge.
Blumenthal, T.D., in press. Short lead interval startle modification. In: Dawson, M.E., Schell, A.M.,
Böhmelt, A.H. (Eds.), Startle Modification: Implications for Neuroscience, Cognitive Science, and
Clinical Science. Cambridge University Press, Cambridge.
Blumenthal, T.D., Creps, C.L., 1994. Normal startle responding in psychosis-prone college students.
Personality Individual Diff. 17, 345–355.
Blumenthal, T.D., Levey, B.J., 1989. Prepulse rise time and startle reflex modification: different effects
for discrete and continuous pre-pulses. Psychophysiology 26, 158 – 165.
Blumenthal, T.D., Gescheider, G.A., 1987. Modification of the acoustic startle reflex by a tactile
prepulse: the effects of stimulus onset synchrony and prepulse intensity. Psychophysiology 24,
320–327.
Blumenthal, T.D., Tolomeo, E.A., 1989. Bidirectional influences of vibrotactile stimuli on modification
of the human acoustic startle reflex. Psychobiology 17, 315 – 322.
Blumenthal, T.D., Schicatano, E.J., Chapman, J.G., Norris, C.M., Ergenzinger, E.R., 1996. Prepulse
effects on magnitude estimation of startle-eliciting stimuli and startle responses. Percep. Psychophys.
58, 73–80.
Böhmelt, A.H., Dawson, M.E., Vanman, E.J., Schell, A.M., 1996. Pre-stimulus modality and effects of
attention on startle eyeblink modification [abstract]. Psychophysiology 33 (Suppl.), S24.
Boelhouwer, A.J.W., Teurlings, R.F.M.A., Brunia, C.H.M., 1991. The effect of an acoustic warning
stimulus upon the electrically elicited blink reflex in humans. Psychophysiology 28, 133 – 139.
Bohlin, G., Graham, F.K., 1977. Cardiac deceleration and reflex blink facilitation. Psychophysiology 14,
423–430.
Bohlin, G., Graham, F.K., Silverstein, L.D., Hackley, S.A., 1981. Cardiac orienting and startle blink
modification in novel and signal situations. Psychophysiology 18, 603 – 611.
Bolino, F., Di-Michele, V., Di-Cicco, L., Manna, V., Daneluzzo, E., Casacchia, M., 1994. Sensorimotor
gating and habituation evoked by electro-cutaneous stimulation in schizophrenia. Biolog. Psychiatry
36, 670–679.
Bonnet, M., Bradley, M.M., Lang, P.J., Requin, J., 1995. Modulation of spinal reflexes: arousal,
pleasure, action. Psychophysiology 32, 367 – 372.
Bowers, D., Bauer, R., Heilman, K., 1993. The nonverbal affect lexicon: theoretical perspectives from
the neuropsychological studies of affect perception. Neuropsychology 2, 733 – 744.
Bradley, M.M., Cuthbert, B.N., Lang, P.J., in press. Affect and the startle reflex. In: Dawson, M.E.,
Schell, A.M., Böhmelt, A.H. (Eds.), Startle Modification: Implications for Neuroscience, Cognitive
Science, and Clinical Science. Cambridge University Press, Cambridge.
Bradley, M.M., Cuthbert, B.N., Lang, P.J., 1990. Startle reflex modification: emotion or attention?.
Psychophysiology 27, 513–522.
Bradley, M.M., Cuthbert, B.N., Lang, P.J., 1991. Startle and emotion: lateral acoustic probes and the
bilateral blink. Psychophysiology 28, 285 – 295.
Bradley, M.M., Cuthbert, B.N., Lang, P.J., 1993a. Pictures as prepulse: attention and emotion in startle
modification. Psychophysiology 30, 541 – 545.
Bradley, M.M., Lang, P.J., Cuthbert, B.N., 1993b. Emotion, novelty, and the startle reflex: habituation
in humans. Behav. Neurosc. 107, 970 – 980.
Bradley, M.M., Cuthbert, B.N., Lang, P.J., 1996. Lateralized startle probes in the study of emotion.
Psychophysiology 33, 156–161.
Braff, D.L., Geyer, M.A., 1990. Sensorimotor gating and schizophrenia: human and animal studies.
Arch. Gen. Psychiatry 47, 181–188.
Braff, D.L., Sacuzzo, D.P., Geyer, M.A., 1991. Information processing dysfunctions in schizophrenia:
studies of visual backward masking, sensorimotor gating, and habituation. In: Steinhauer, S.R.,
Gruzelier, J.H., Zubin, J. (Eds.), Handbook of Schizophrenia, vol. 5, Neuropsychology, Psychophysiology and Information Processing. Elsevier, Amsterdam, pp. 303 – 334.
Braff, D.L., Grillon, C., Geyer, M.A., 1992. Gating and habituation of the startle reflex in schizophrenic
patients. Arch. Gen. Psychiatry 49, 206 – 215.
38
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
Braff, D.L., Swerdlow, N.R., Geyer, M.A., 1995. Gating and habituation deficits in the schizophrenia
disorders. Clin. Neurosci. 3, 131–139.
Braff, D., Stone, C., Callaway, E., Geyer, M., Glick, I., Bali, L., 1978. Prestimulus effects on human
startle reflex in normals and schizophrenics. Psychophysiology 15, 339 – 343.
Brown, J.S., Kalish, H.I., Farber, I.E., 1951. Conditioned fear as revealed by magnitude of startle
response to an auditory stimulus. J. Exp. Psychol. 32, 317 – 328.
Butler, R.W., Jenkins, M.A., Braff, D.L., 1993. On the abnormality of normal comparison groups: the
identification of psychosis-proneness and substance abuse in putative normal research subjects. Am.
J. Psych. 150, 1386–1391.
Cadenhead, K.S., Braff, D.L., in press. Schizophrenia spectrum disorders. In: Dawson, M.E., Schell,
A.M., Böhmelt, A.H. (Eds.), Startle Modification: Implications for Neuroscience, Cognitive Science,
and Clinical Science. Cambridge University Press, Cambridge.
Cadenhead, K.S., Braff, D.L., 1992. Which criteria select ‘psychosis-prone’ individuals?. Biolog. Psychiatry 31, 16lA–162A.
Cadenhead, K.S., Geyer, M.A., Braff, D.L., 1993. Impaired startle prepulse inhibition and habituation
in patients with schizotypal personality disorder. Am. J. Psych. 150, 1862 – 1867.
Center for the Study of Emotion and Attention [CSEA-NIMH], 1995. The international affective picture
system: digitized photographs. Gainsville, FL: The Center for Research in Psychophysiology,
University of Florida.
Chapman, L.J., Chapman, J.P., 1987. The search for symptoms predictive of schizophrenia. Schizophr.
Bull. 13, 497–503.
Chapman, L.J., Chapman, J.P., Kwapil, T.R., Eckblad, M., Zinser, M.C., 1994. Putatively psychosisprone subjects 10 years later. J. Ab. Psychol. 103, 171 – 183.
Clarkson, M.G., Berg, W.K., 1984. Bioelectric and potentiometric measures of eyeblink amplitude in
reflex modification paradigms. Psychophysiology 21, 237 – 241.
Cleckley, H. 1976. The Mask of Sanity, fifth ed. Mosby, St. Louis, MO.
Cohen, L.H., Hoffman, H.S., Stitt, C.L., 1981. Sensory magnitude estimation in the context of reflex
modification. J. Exp. Psychol.: Human Perception and Performance 7, 1363 – 1370.
Cook, E.W., in press. Affective individual differences, psychopathology, and startle reflex modification.
In: Dawson, M.E., Schell, A.M., Böhmelt, A.H. (Eds.), Startle Modification: Implications for
Neuroscience, Cognitive Science, and Clinical Science. Cambridge University Press, Cambridge.
Cook, E.W., Davis, T.L., Hawk, L.W., Spence, E.L., Gautier, C.H., 1992. Fearfulness and startle
potentiation during aversive visual stimuli. Psychophysiology 29, 633 – 645.
Cook, E.W., Hawk, L. Jr., Davis, T.L., Stevenson, V.E., 1991. Affective individual differences and
startle reflex modulation. J. Ab. Psychol. 100, 5 – 13.
Cuthbert, B.N., Bradley, M.M., Lang, P.J., 1996. Probing picture perception: activation and emotion.
Psychophysiology 33, 103–111.
Davis, M. 1996. Fear-potentiated startle in the study of animal and human emotions. In: Kavanaugh,
R.D., Zimmerberg, B., Fein, S. (Eds.), Emotion: Interdisciplinary perspectives. Erlbaum, Hillsdale,
NJ, pp. 61–89.
Davis, M., Astraehan, B., 1978. Conditioned fear and startle magnitude: effects of different footshock
or backshock intensities. J. Exp. Psychol. 4, 95 – 103.
Davis, M., Gendelman, D., Tischler, M., Gendelman, P., 1982. A primary acoustic startle circuit: lesion
and stimulation studies. J. Neurosci. 2, 791 – 805.
Davis, M., Schlesinger, L.S., Sorenson, C.A., 1989. Temporal specificity of fear conditioning: effects of
different conditioned stimulus-intervals on the fear-potentiated startle effect. J. Exp. Psychol. 15,
295–310.
Davis, M., Walker, D.C., Lee, Y., in press. Neurophysiology and neuropharmacology of startle and its
affective modification. In: Dawson, M.E., Schell, A.M., Böhmelt, A.H. (Eds.), Startle Modification:
Implications for Neuroscience, Cognitive Science, and Clinical Science. Cambridge University Press,
Cambridge.
Dawson, M.E., 1990. Psychophysiology at the interface of clinical science, cognitive science, and
neuroscience. Psychophysiology 27, 243 – 255.
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
39
Dawson, M.E., Filion, D.L., Schell, A.M., 1989. Is the elicitation of the orienting response associated
with the allocation of resources?. Psychophysiology 26, 560 – 572.
Dawson, M.E., Hazlett, E.A., Filion, D.L., Nuechterlein, K.H., Schell, A.M., 1993. Attention and
schizophrenia: impaired modulation of the startle reflex. J. Abnormal Psychol. 102, 633 – 641.
Dawson, M.E., Hazlett, E.A., Filion, D.L., Neuchterlein, K.H., Schell, A.M., 1995. Attention, startle
eye-blink modification, and psychosis-proneness. In: Raine, A., Lenez, T., Medniek, S.A. (Eds.),
Schizotypal Personality Disorder. Cambridge University Press, Cambridge.
Dawson, M.E., Schell, A.M., Böhmelt, A.H., in press. Startle Modification: Implications for Neuroscience, Cognitive Science, and Clinical Science. Cambridge University Press, Cambridge.
Dawson, M.E., Schell, A.M., Swerdlow, N.R., Filion, D.L., 1997. Cognitive, clinical, and neurophysiological implications of startle modification. In: Lang, P.J., Simons, R.F., Balaban, M. (Eds.),
Attention and Orienting: Sensory and Motivational Processes. Erlbaum, Hillsdale, NJ, pp. 257 – 327.
de Jong, P.J., Merckelbach, H., Arntz, A., 1991. Eyeblink startle response in spider phobics before and
after treatment: a pilot study. J. Psychopathol. Behavior. Assessment 3, 213 – 223.
de Jong, P.J., Arntz, A., Merckelbach, H., 1993. The startle probe response as an instrument for
evaluating exposure effects in spider phobia. Special Issue: Phobia: etiological, cognitive, and
physiological aspects. Adv. Behav. Res. Therapy 15, 301 – 316.
DelPezzo, E.M., Hoffman, H.S., 1980. Attention factors in the inhibition of a reflex by a visual stimulus.
Science 2 (10), 673–674.
Ehrlichman, H., Brown, S., Zhu, J., Warrenburg, S., 1995. Startle reflex modulation during exposure to
pleasant and unpleasant odors. Psychophysiology 32, 150 – 154.
Filion, D.L., Ciranni, M., 1994. The functional significance of prepulse inhibition: A test of the
protection of processing theory [abstract]. Psychophysiology 31 (Suppl.), S46.
Filion, D.L., Seljos, K.A., Hazlett, E.A., in press. Behavioral analogies of short lead interval startle
inhibition. In: Dawson, M.E., Schell, A.M., Böhmelt, A.H. (Eds.), Startle Modification: Implications
for Neuroscience, Cognitive Science, and Clinical Science. Cambridge University Press, Cambridge.
Filion, D.L., Dawson, M.E., Schell, A.M., 1993. Modification of the acoustic startle reflex eyeblink: A
tool for investigating early and late attentional processes. Biolog. Psychol. 35, 185 – 200.
Filion, D.L., Dawson, M.E., Schell, A.M., 1994. Probing the orienting response with startle modification
and secondary reaction time. Psychophysiology 31, 68 – 78.
Filion, D.L., Dawson, M.E., Schell, A.M., Hazlett, E.A., 1991. The relationship between skin conductance orienting and the allocation of processing resources. Psychophysiology 28, 410 – 425.
Flaten, M.A., Blumenthal, T.D., 1996. Effects of positive and negative stimulus onset asynchronous
between weak stimuli and blink reflex eliciting stimuli. J. Psychophysiol. 0, 189 – 197.
Ford, J.M., Roth, W.T. in press. Modification of reflexively elicited event related potential (ERP)
components by leading stimuli. In: Dawson, M.E., Schell,, A.M. Böhmelt, A.H. (Eds.), Startle
Modification: Implications for Neuroscience, Cognitive Science, and Clinical Science. Cambridge
University Press, Cambridge.
Foss, J.A., Ison, J.R., Torre, J. Jr., Wansaek, S., 1989. The acoustic startle response and disruption of
aiming: II. Modulation by fore-warning and preliminary stimuli. Human Factors 3 (1), 319 – 333.
Freedman, R., Adler, L.E., Gerhardt, G.A., Waldo, M., Baker, N., Rose, G.M., Drebing, C.,
Nagamoto, H., Bickford-Wimer, P., Franks, R., 1987. Neurobiological studies of sensory gating in
schizophrenia. Schizophr. Bull. 13, 669 – 678.
Geyer, M.A., Swerdlow, N.R., Mansbach, R.S., Braff, D.L., 1990. Startle response models of sensorimotor gating and habituation deficits in schizophrenia. Brain Res. Bull. 25, 485 – 498.
Graham, F.K., 1975. Presidential Address, 1974. The more or less startling effects of weak pre-stimulation. Psychophysiology 12, 238–248.
Graham, F.K., 1979. Distinguishing among orienting, defence, and startle reflexes. In: Kimmel, H.D.,
van Olst, E.H., Orlebeke, J.F. (Eds.), The Orienting; Reflex in Humans. Erlbaum, Hillsdale, NJ, pp.
137–167.
Graham, F.K., 1980. Control of reflex blink excitability. In: Thompson, R.F., Hieks, L.H., Shvyrkov,
V.B. (Eds.), Neural Mechanisms of Goal-Directed Behaviour and Learning. Academic Press, New
York, pp. 511–519.
40
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
Graham, F.K., 1992. Attention: The heartbeat, the blink, and the brain. In: Campbell, B.A., Hayne, H.,
Richardson, R. (Eds.), Attention and Information Processing in Infants and Adults: Perspectives
from Human and Animal Research. Erlbaum, Hillsdale, NJ, pp. 3 – 29.
Graham, F.K., Murray, G.M., 1977. Discordant effects of weak pre-stimulation on magnitude and
latency of the reflex blink. Physiolog. Psychol. 5, 108 – 114.
Graham, F.K., Putnam, L.E., Leavitt, L.A., 1975. Lead-stimulation effects of human cardiac orienting
and blink reflexes. J. Exp. Psychol.: Human Perception and performance 104, 175 – 182.
Graham, F.K., Strock, B.D., Zeigler, B.L., 1981. Excitatory and inhibitory influences on reflex
responsiveness. In: Collins, W.A. (Ed.), Minnesota Symposium on Child Psychology 14, Erlbaum,
Hillsdale, NJ, pp. 1–38.
Grillon, C., Ameli, R., Chamey, D.S., Krystal, J., Braff, D., 1992. Startle gating deficits occur across
prepulse intensities in schizophrenic patients. Biolog. Psychiat. 32, 939 – 943.
Grillon, C., Ameli, R., Merikangas, K., Woods, S.W., Davis, M., 1993. Measuring the time course of
anticipatory anxiety using the fear-potentiated startle reflex. Psychophysiology 30, 340 – 346.
Grillon, C., Ameli, R., Woods, S.W., Merikangas, K., Davis, M., 1991. Fear-potentiated startle in
humans: effects of anticipatory anxiety on the acoustic blink reflex. Psychophysiology 28, 588 – 595.
Hackley, S.A. in press. Blink reflex studies of attention, consciousness, and automaticity. In: Dawson,
M.E., Schell, A.M., Böhmelt, A.H. (Eds.), Startle Modification: Implications for Neuroscience,
Cognitive Science, and Clinical Science. Cambridge University Press, Cambridge.
Hackley, S.A., Graham, F.K., 1983. Early selective attention effects on cutaneous and acoustic blink
reflexes. Physiolog. Psychol. 11, 235– 242.
Haekley, S.A., Graham, F.K., 1987. Effects of attending selectively to the spatial position of reflex-eliciting and reflex-modulating stimuli. J. Exp. Psychol.: Human Perception and Performance 13,
411–424.
Haerich, P., 1994. Startle reflex modification: effects of attention vary with emotional valence. Psycholog. Sci. 5, 407–410.
Hamm, A.O., Cuthbert, B.N., Globisch, J., Vaitl, D., 1997. Fear and the startle reflex: blink modulation
and autonomic response patterns in animal and mutilation fearful subjects. Psychophysiology 34,
97–107.
Hamm, A.O., Greenwald, M.K., Bradley, M.M., Lang, P.J., 1993. Emotional learning, hedonic change,
and the startle probe. J. Abn. Psychol. 102, 453 – 465.
Harbin, T.J., Berg, W.K., 1986. The effects of age and attention upon reflex inhibition. Biolog. Psychol.
22, 81–94.
Hawk, L.W., Cook, E.W., 1997. Affective modulation of tactile startle. Psychophysiology 34, 23 – 31.
Hazlett, E.A., Buchsbaum, M.S., Hazuedar, M.M., Gemmans, M.K., Luu, C., 1995. Cortical-striatalthalamic circuit analysis from PET during a startle blink paradigm [abstract]. Psychophysiology 32
(Suppl.), S39.
Hazlett, E.A., Dawson, M.E., Nuechterlein, K.H., Filion, D.L., 1993. Startle modification during a
continuous performance task: Implications for schizophrenia [abstract]. Psychophysiology 30
(Suppl.), S33.
Hoffman, H.S., Ison, J.R., 1980. Reflex modification in the domain of startle: I. Some empirical findings
and their implications for how the nervous system processes sensory input. Psycholog. Rev. 87,
175–189.
Hoffman, H.S., Wible, B.L., 1969. Temporal parameters. In startle facilitation by steady background
signals. J. Acoust. Soc. Am. 45, 7–12.
Ison, J.R., Ashkenazi, B., 1980. Effects of a warning stimulus on reflex elicitation and reflex inhibition.
Psychophysiology 17, 586–591.
Ison, J.R., Hammond, G.R., 1971. Modifications of the startle reflex in the rat by changes in the
auditory and visual environments. J. Comparative Physiolog. Psychol. 75, 418 – 430.
Ison, J.R., Hoffman, H.S., 1983. Reflex modification in the domain of startle: II. The anomalous history
of a robust and ubiquitous phenomenon. Psychol. Bull. 94, 3 – 17.
Ison, J.R., O’Conner, K., Bowen, G.P., Bocirnea, A., 1991. Temporal resolution of gaps in noise by the
rat is lost with functional decortication. Behav. Neurosci. 105, 33 – 40.
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
41
Jensen, D.M., Frijda, N.H., 1994. Modulation of the acoustic startle response by film induced fear and
sexual arousal. Psychophysiology 31, 565 – 571.
Jennings, P., Schell, A.M., Filion, D.L., Dawson, M.E., 1996. Tracking early and late attentional
processes with startle modification. Psychophysiology 33, 148 – 155.
Koch, M., Lingenhohl, K., Pilz, P.K.D., 1992. Loss of the acoustic startle response following neurotoxic
lesions of the caudal pontine reticular formation: possible role of giant neurons. Neuroscience 49,
617–625.
Kring, A.M., Neale, J.M., 1996. Do schizophrenic patients show a disjunctive relationship among
expressive, experiential, and psychophysiological components of emotion?. J. Abn. Psychol. 105,
249–257.
Lang, P.J., 1995. The emotion probe: studies of motivation and attention. Am. Psychol. 50, 372 – 385.
Lang, P.J., Bradley, M.M., Cuthbert, B.N., 1990. Emotion, attention, and the startle reflex. Psycholog.
Rev. 97, 377–395.
Lang, P.J., Bradley, M.M., Cuthbert, B.N., 1992. A motivational analysis of emotion: reflex-cortex
connections. Psycholog. Sci. 3, 44–49.
Lang, P.J., Bradley, M.M., Cuthbert, B.N., 1995. International affective picture system (IAPS):
Technical manual and affective ratings. Gainsville, FL. The Center for Research in Psychophysiology, University of Florida.
Lang, P.J., Bradley, M.M., Cuthbert, B.N., Patrick, C.J., 1993. Emotion and psychopathology: a startle
probe analysis. Prog. Exp. Personal. Psychopathol. Res. 16, 163 – 199.
Lazarus, R.S., 1984. On the primacy of cognition. Am. Psychol. 39, 1214 – 1229.
Lee, Y., Lopez, D.E., Meloni, E., Davis, M., 1996. A primary acoustic startle pathway: obligatory role
of cochlear root neurons and the reticular pontis caudalis. J. Neurosci. 6, 3775 – 3789.
Leitner, D.S., Cohen, M.E., 1985. Role of the inferior colliculus in the inhibition of acoustic startle in
the rat. Physiol. Behav. 34, 65–70.
Lipp, O.V., Siddle, D.A.T. in press. Startle modification during orienting and Pavlovian conditioning.
In: Dawson, M.E., Schell, A.M., Böhmelt, A.H. (Eds.), Startle Modification: Implications for
Neuroscience, Cognitive Science, and Clinical Science. Cambridge University Press, Cambridge.
Lipp O.V., Siddle, D.A.T., Arnold, S.L., 1994. Psychosis proneness in a non-clinical sample II: A
multi-experimental study of ‘attentional malfunctioning.’ Personality and Individual Differences, 17,
405–424.
Lipp, O.V., Siddle, D.A.T., Dall, P.J., 1997. The effect of foreground stimulus modality on blink
magnitude startle modulation. Psychophysiology 34, 340 – 347.
Lukoff, D., Nuechterlein, K.H., Ventura, J., 1986. Appendix A: manual for expanded Brief Psychiatric
Rating Scale (BPRS). Schizophr. Bull. 12, 594 – 602.
McDowd, J.M., Filion, D.L., Harris, M.J., Braff, D.L., 1993. Sensory gating and inhibitory function in
late-life schizophrenia. Schizophr. Bull. 19, 733 – 746.
Miltner, W., Matjak, M., Braun, C., Diekmann, H., Brody, S., 1994. Emotional qualities of odors and
their influence on the startle reflex in humans. Psychophysiology 31, 107 – 110.
Morgan, C.A., Grillon, C., Southwick, S.M., Davis, M., Charney, D.S., 1995. Fear potentiated startle
in post-traumatic stress disorder. Biolog. Psychiatry 38, 378 – 385.
Norris, C.M., Blumenthal, T.D., 1995. Evidence for the protection of preattentive processing during
inhibition of the acoustic startle response [abstract]. Psychophysiology (Suppl)., 32, 57.
Nuechterlein, K.H., Dawson, M.E., Gitin, M., Ventura, J., Goldstein, M.J., Snyder, K.S., Yee, C.M.,
Mintz, J., 1992. Developmental processes in schizophrenic disorders: Longitudinal studies of
vulnerability and stress. Schizophr. Bull. 18, 387 – 425.
Ornitz, E.M. in press. Developmental aspects of startle modification. In: Dawson, M.E., Schell, A.M.,
Böhmelt, A.H. (Eds.), Startle Modification: Implications for Neuroscience, Cognitive Science, and
Clinical Science. Cambridge University Press, Cambridge.
Ornitz, E.M., Guthrie, D., Kaplan, A.R., Lane, S.J., Nomman, R.J., 1986. Maturation of startle
modulation. Psychophysiology 23, 624 – 634.
Ornitz, E.M., Guthrie, D., Sadeghpour, M., Sugiyama, T., 1991. Maturation of pre-stimulation-induced
startle modulation in girls. Psychophysiology 28, 11 – 20.
42
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
Ornitz, E.M., Hanna, G.L., de-Traversay, J., 1992. Pre-stimulation-induced startle modulation in
attention-deficit hyperactivity disorder and nocturnal enuresis. Psychophysiology 29, 437 – 451.
Overall, J.E., Gorham, D.R., 1962. The Brief Psychiatric Rating Scale. Psycholog. Rep. 10, 799 – 812.
Patrick, C.J. 1994. Emotion and psychopathy: startling new insights. Psychophysiology 319 – 330.
Patrick, C.J., Berthot, B.D., Moore, J.D., 1996. Diazopam blocks fear-potentiated startle in humans. J.
Abn. Psychol. 105, 89–96.
Patrick, C.J., Bradley, M.M., Lang, P.J., 1993. Emotion in the criminal psychopath: startle reflex
modulation. J. Abn. Psychol. 102, 82 – 92.
Patrick, C.J., Lang, A.R. (in press). Psychopathic traits and intoxicated states: Affective concomitants
and conceptual links. In: M.E. Dawson, A.M. Schell, A.H. Böhmelt (Eds.), Startle Modification:
Implications for Neuroscience, Cognitive Science, and Clinical Science. Cambridge University Press,
Cambridge.
Perlstein, W.M., Fiorito, E., Simons, R.F., Graham, F.K., 1989. Pre-stimulation effects on reflex blink
and evoked potentials in normal and schizotypal subjects [abstract]. Psychophysiology 26 (Suppl.),
S48.
Perlstein, W.M., Fiorito, E., Simons, R.F., Graham, F.K., 1993. Lead stimulation effects on reflex blink,
exogenous brain potentials, and loudness judgments. Psychophysiology 30, 347 – 358.
Perry, W., Braff, D.L., 1994. Information-processing deficits and thought disorder in schizophrenia. Am.
J. Psychiatry 151, 363–367.
Putnam, L.W., in press. Long lead interval startle modification. In: Dawson, M.E., Schell, A.M.,
Böhmelt, A.H. (Eds.), Startle Modification: Implications for Neuroscience, Cognitive Science, and
Clinical Science. Cambridge University Press, Cambridge.
Putnam, L.E. 1990. Great expectations: anticipatory responses of the heart and brain. In: Rohrbaugh,
J.W., Parasuraman, R., Johnson, R., Jr. (Eds.), Event-Related Brain Potentials: Basic Issues and
Applications. Oxford University Press, New York, pp. 109 – 129.
Robinson, J.D., Vrana, S.R., 1995. Modulation of the startle prepulse effect during emotional imagery
[abstract]. Psychophysiology 32 (Suppl.), S64.
Schell, A.M., Dawson, M.E., Hazlett, E.A., Filion, D.L., 1995. Attentional modulation of startle in
psychosis-prone college students. Psychophysiology 32, 266 – 273.
Schlenker, R., Cohen, R., Hopmann, G., 1995. Affective modulation of the startle reflex in schizophrenic
patients. Eur. Arch. Psychiatry Clin. Neurosci. 245, 309 – 318.
Schwarzkopf, S.B., Lamberti, J.S., Smith, D.A., 1993. Concurrent assessment of acoustic startle and
auditory P50 evoked potential measures of sensory inhibition. Biol. Psychiatry 815 – 828.
Sechenov, I.M. 1965. Reflexes in the brain. (S. Belsky, trans.) MIT Press, Cambridge, Mass. (Originally
published in 1863, St. Petersburg: Sushchinski).
Seljos, K.A., Dawson, M.E., Schell, A.M., 1995. Availability of information processing resources affects
startle eyeblink and prepulse inhibition [abstract]. Psychophysiology 32 (Suppl.), S68.
Seljos, K.A., Filion, D.L., Dawson, M.E., Schell, A.M., 1994. Background noise disrupts attentional
modulation of the startle reflex [abstract]. Psychophysiology 31 (Suppl.), S90.
Siegel, A., 1967. Stimulus generalization of a classically conditioned response along a temporal
dimension. J. Compar. Physiolog. Psychol. 64, 461 – 466.
Silverstein, L.D., Graham, F.K., Bohlin, G., 1981. Selective attention effects on the reflex blink.
Psychophysiology 18, 240–247.
Silverstein, L.D., Graham, F.K., Calloway, J.M., 1980. Preconditioning and excitability of the human
orbicularis oculi reflex as a function of state. Electroencephalogr. Clin. Neurophysiol. 48, 406 – 417.
Simons, R.F., Giardina, B.D., 1992. Reflex modification in psychosis-prone young adults. Psychophysiology 29, 8–16.
Simons, R.F., Zelson, M.F., 1985. Engaging visual stimuli and reflex blink modification. Psychophysiology 22, 44–49.
Sugawara, M., Sadeghpour, M., De-Traversay, J., Ornitz, E.M., 1994. Prestimulation-induced modulation of the P300 component of event related potentials accompanying startle in children. Electroencephalogr. Clin. Neurophysiol. 90, 201 – 213.
Stritzke, W.G.K., Patrick, C.J., Lang, A.R., 1995. Alcohol and human emotion: a multidimesional
analysis incorporating startle-probe methodology. J. Abn Psychol. 04, 114 – 122.
D.L. Filion et al. / Biological Psychology 47 (1998) 1–43
43
Swerdlow, N.R., Benbow, C.H., Zisook, S., Geyer, M.A., Braff, D.L., 1993. A preliminary assessment
of sensorimotor gating in patients with obsessive compulsive disorder. Biol. Psychiatry 33, 298 – 301.
Swerdlow, N.R., Caine, S.B., Braff, D.L., Geyer, M.A., 1992. Neural substrates of sensorimotor gating
of the startle reflex: a review of recent findings and their implications. J. Psychopharmocol. 6,
176–190.
Swerdlow, N.R., et al., 1995a. ‘Normal’ personality correlates of sensorimotor, cognitive, and visuospatial gating. Biol. Psychiatry 37, 286–299.
Swerdlow, N.R., Geyer, M.A. in press. Neurophysiology and neuropharmacology of short lead interval
startle modification. In: Dawson, M.E., Schell, A.M., Böhmelt, A.H. (Eds.), Startle Modification:
Implications for Neuroscience, Cognitive Science, and Clinical Science. Cambridge University Press,
Cambridge.
Swerdlow, N.R., Paulsen, J., Braff, D.L., Butters, N., Geyer, M.A., Swenson, M.R., 1995b. Impaired
prepulse inhibition of acoustic and tactile startle response in patients with Huntington’s disease. J.
Neurol. Neurosurg. Psychiatry 58, 192 – 200.
Vanman, E.J., Böhmelt, A.H., Dawson, M.E., Schell, A.M., 1996. The varying time courses of
attentional and affective modulation of the startle eyeblink reflex. Psychophysiology 33, 691 – 698.
Vrana, S.R., 1994. Startle reflex response during sensory modality specific disgust, anger, and neutral
imagery. J. Psychophysiol. 8, 211–218.
Vrana, S.R., Lang, P.J., 1990. Fear imagery and the startle-probe reflex. J. Abn. Psychol. 99, 189 – 197.
Vrana, S.R., et al., 1992. Startle reflex modification as an outcome measure in the treatment of phobia:
two case studies. Behav. Assessment 14, 279 – 291.
Vrana, S.R., et al., 1988. The startle probe response: A new measure of emotion?. J. Abn. Psychol. 97,
487–491.
Witvliet, C.V., Vrana, S.R., 1995. Psychophysiological responses as indices of affective dimensions.
Psychophysiology 32, 436–443.
Wu, M.F., et al., 1984. Startle reflex inhibition in the rat: its persistence after extended repetition of the
inhibitory stimulus. J. Exp. Psychol. 10, 221 – 228.
Wynn, J.K., Schell, A.M., Dawson, M.E., 1996. Effects of prehabituation of the prepulse on startle
eyeblink modification [abstract]. Psychophysiology 33 (Suppl.), S92.
Zajonc, R.B., 1980. Feeling and thing: Preferences need no inferences. Am. Psychol. 35, 151 – 175.
Zelson, M.F., Simons, R.F., 1986. Sustained attention in Type A and Type B subjects: a blink reflex
analysis. Psychophysiology 23, 385–392.
.