Fear of Wolf and Bear – physiological responses and negative association. Anders Flykt1, Maria Johansson2, Jens Karlsson3, & Sofie Lindeberg2 1 Academy of Health and Occupational Studies, Department of Social Work and Psychology, University of Gävle, Sweden 2 Environmental Psychology, Deptartment of Architecture and Built Environment, Lund University, Sweden 3 Grimsö Wildlife Research Station, Swedish University of Agricultural Sciences, Sweden Corresponding author: Anders Flykt Department of Social Work and Psychology, Academy of Health and Occupational studies University of Gävle S-801 76 Gävle Sweden Phone: +46-26-648574 E-mail: Anders.Flykt@hig.se / ansflt@hig.se Key words: Fear, Bear, Wolf, IAT, Visual search, SCR, Eye movements, HR 1 Abstract Participants were fearful of bears, of both bear and wolf, and not fearful of bears or wolves. Firstly pictures of bears, wolves, moose, and hares were displayed, eyemovements, skin conductance, and ECG were recorded. Secondly participants decided if a hare picture was present among moose pictures where a picture of a wolf or a bear could occur. Thirdly, bear, wolf, and hare pictures were sorted with good or bad words. Independently of fearfulness, bear pictures showed stronger physiological responding and wolf pictures showed stronger negative association. The bear fear only group showed somewhat stronger physiological responding to bears while the bear and wolf fearful group showed more difficulty in associating bears with good words. When a feared animal occurred during the search task, response time was prolonged. Fear of bears and wolves seem to be driven by elaborate cognitive processing rather than by specific fear circuits. 2 Fear of Wolf and Bear – physiological responses and negative association. The probability of humans being attacked by large carnivores in Sweden is at present relatively small (Linnell et al., 2001), but in a recent survey 44 percent of the individuals reported that they were afraid of encountering brown bears in the forest and 25 percent reporting that they were afraid of encountering wolves (Ericsson, Sandström, Kindberg, & Stoen, 2010). The figures for people with experience of living in areas with wolves and bears are similar. Forty-two percent of those in areas with presence of brown bear report fear of encountering brown bear in the forest and 33 percent of those in areas with presence of wolf report fear of encountering wolf in the forest (Johansson, Karlsson, Pedersen, & Flykt, 2012). When asked about how fearful they are of attacks of brown bear respectively wolf, only 30% in both cases said that they feel no fear what so ever. As fear of wolves and bears is so widespread it affects a large proportion of humans living in areas with large carnivores and thereby also has a potentially large impact on decision makers and large carnivore policy. In order to understand and eventually develop measures to meet human fear of large carnivores it is essential to know more about the mental processes involved. Up to date no such attempts have been made regarding fear for large carnivores. Fear of snakes and spiders has been intensively studied and may provide a useful theoretical framework and relevant comparison. Despite that there is basically no environmental constraints for everyday life caused by snakes or spiders, a significant proportion (12 % in females and 3% in males) of the Swedish adult population is fearful of these animals (see Fredrikson, Annas, Fischer, &Wik, 1996). Sweden has three species of snakes, adder (Vipera Berus), grass- 3 snake (Natrix natrix), and smooth snake (Coronella austriaca), and it is only the adder that is potentially lethal for individuals with allergic responses to adder poison. Sweden has no harmful spiders. From an evolutionary theory view point, the explanation could be that ancestors of humankind would have had a reproductive advantage by acquiring fear of snakes and spiders fast, without effort, and resistant to extinction (i.e. the preparedness theory, see Seligman, 1971), it has also been proposed that the visual system in primates partly has evolved with respect to the presence of snakes (Isbell, 2006). Thus, snakes and spiders should posses a very specific place in primate information processing, and may therefore explain the large proportion of reported fear of snakes and spiders. Furthermore, Seligman (1970, 1971) suggested evolutionary old species, like insects, rats, reptiles, and birds, to be overrepresented in phobic or subclinical fear. Thus, human fear for mammals like brown bear and wolf would not be based on the same evolutionary old processing. Mental processing associated with fear for snakes and spiders have been intensively researched. Some of the processes addressed have been attentional capture of the feared animal, (Öhman, Flykt, & Esteves, 2001), dwell time on the feared animal (Miltner, Krieschel, Hecht, Trippe, & Weiss, 2004; Gerdes, Pauli, & Alpers, 2009) automaticity of responding to the feared animal (Öhman & Soares, 1994) and behavioral influences of the feared animal (Sabatinelli, Bradley, & Lang, 2001, see also Flykt & Caldara, 2006). In order to test for differences between fear for large carnivores and fear of snakes and spiders, we investigated fear of bear and fear of wolf using experimental paradigms previously employed in research investigating psychological processing in participants fearful of snakes and spiders. As the studies investigating fear for snakes and spider often have used participants fearful of snakes, but not spiders, and the reversed, we 4 intended to select participants fearful of bears, but not fearful of wolf (BF), the reversed, and participants not fearful of neither bears nor wolves (NF). However, did not find enough individuals fearful of wolves only, instead we managed to recruit individuals fearful of both wolves and bears (BWF). Reaction times, eye movement data, heart rate change scores, and skin conductance responses (SCRs) were used as dependent measures. Picture viewing Several studies investigating the responses to snakes and spiders in snake or spider fearful participants have presented the feared animal on a screen for a brief period of time. The responses evoked by the presentation have been measured and compared with the presentation of pictures of not feared animals or other stimuli. For example Öhman and Soares (1993, 1994) recruited participants fearful of either snakes or spiders and presented them with pictures of snakes, spiders, flowers, and mushrooms. The common finding was larger SCRs to pictures of the feared animal than other stimuli. A similar result was obtained for heart rate change scores in a study with spider fearful, but not snake fearful, women (Flykt & Bjärtå, 2008). In that study spider pictures was contrasted against five other animal pictures, namely; beetle, snake, turtle, wolf, and rabbit. The pictures presented in that study covered the full screen and the picture display in the Öhman and Soares (1993, 1994) studies had a similar size of stimulus and distance to participant ratio. The perception of the physical closeness to the threat is related to the physiological response (Lang, Bradley, & Cuthbert, 1997) according to the predator imminence continuum (see Fanselow & Lester, 1988). The participants in the Flykt and Bjärtå study showed less deceleration 5 and more acceleration in heart rate changes for spider pictures than the other animal pictures. This effect occurred even though the participants also had tasks to handle during the picture viewing. This is important as the vigilance imposed by the task might be a critical factor for the eliciting of physiological responding. It has been suggested, based on eye tracking results, that when spiders serve as task-irrelevant-distractors to spider fearful participants the participants dwell on these stimuli longer than other task-irrelevant-distractors (Gerdes, Alpers, & Pauli, 2008). Moreover, Gerdes, Pauli, and Alpers (2009) found that spiders are attended to faster than other stimuli tested, but that this was a general effect not related to if the participants were spider fearful or not. However, to our knowledge no study has looked at gaze direction to a feared animal when that is the only presented item on which it is explicitly stated that the participants should focus on. With the rationale that a person can avoid or attend more reliably to a stimulus when there is no alternative stimuli present (that may distract), it is expected that an effect would depend on the feared animal. Visual search Eye-movement is a complement to other dependent measures in visual search tasks. Eyemovements give information about where the participant is looking and for how long, it also provides an opportunity to see if the overt attention is dissociated from reaction times (as shown by Derakshan & Koster, 2011) which are supposed to mirror the detection time. If snakes and spiders are phylogenic old threats in the evolution of humans, conscious awareness should not be a necessity for the elicitation of 6 corresponding fear responses. Öhman and Soares (1993, 1994) masked pictures of the feared animal and other stimuli to prevent conscious recognition of the different stimuli, while measuring SCRs. These experiments showed an elevated SCR to the masked pictures of the feared animal as compared to other stimuli. This has been interpreted in terms of automaticity, assisting the conscious controlled processing in the detection of these potential threats (Öhman & Mineka, 2001). This idea was empirically supported in a visual search study by Öhman, Flykt, and Esteves (2001). Participants fearful of either snakes or spiders had the task to detect a deviant stimulus in a search array. The deviant stimulus could be a snake or a spider among a background of flowers or mushrooms or a flower or mushroom among a background of snakes and spiders. The results showed shorter reaction times (RTs) for the feared animal than other deviant pictures, suggesting a faster detection of the feared stimuli. In another visual search study with participants fearful or either snakes or spiders Flykt and Caldara (2006) showed larger electrical activity over the scalp for a deviant feared animal than a not feared animal at a time window of 500-700 ms after stimulus onset. Flykt (2006) argued that the shorter RTs to the feared animal might be a consequence of a motor preparation for handling the threat rather than an effect of early detection. Independent of the origin of this effect, it seems relatively stable. To address the question if the feared animal captures attention or attention dwells on the feared animal Miltner, Krieschel, Hecht, Trippe, and Weiss (2004; See also Lipp & Waters, 2007) used a modified version of the visual search paradigm with their spider fearful participants. In one condition the participants had to find a mushroom target among 15 flower pictures and in another the participants had to find a mushroom target 7 among 14 flower picture background and one spider picture (we call this an oddball distractor). The RTs to the mushroom target for those search arrays that also had a spider picture in the array were longer than for those without the spider picture. This result suggests that the spider picture was taking processing resources from the target stimuli. However, in the Miltner, et al. (2004) study, also eye moments were measured to investigate the overt attention showing longer scan paths (i.e. time) for a mushroom target with a spider oddball distractor in the search array than for a mushroom target without this distractor. This delay supports a longer dwelling on the threat rather an attentional capture, as was suggested by Öhman et al. (2001). Similar findings and arguments have been made by Gerdes et al. (2008). Independent of whether the prolongation of RTs to feared task irrelevant distractor is due to an automatic capturing (Öhman et al., 2001) or a slowed down disengagement process (Miltner et al., 2004), or a decrease of inhibition of fear evoking distractors and a reduction of resources available for shifting attentional focus to the task at hand (Eysenck, Derakshan, Santos, & Calvo, 2007) the visual search results by Miltner et al., (2004) are indicative of fear responding to a supposedly feared animal. Moreover, it serves as a measure of performance loss due to exposure to the feared animal. Implicit association test (IAT) The implicit association test is based on the assumption that there is strong response compatibility for pictures, words, and concepts that all are categorized as for example Good or Bad (see e.g. Greenwald, McGhee, & Schwartz, 1998). Thus, when a word and a picture are sorted into the same category (e.g. both good or both bad) the 8 requested response will be faster and higher in accuracy compared to a word and a picture categorized into different categories (e.g. good and bad). Teachman and coworkers (Teachman, Gregg, & Woody, 2001) showed that in IAT with spider or snake fearful participants, those fearful of snakes had longer RTs for responding to snake – good and spider – bad, than for the reversed condition, and that the opposite was true for those fearful of spiders. In another IAT study by Teachman and Woody (2003) the spider fearful participants showed longer RTs for the combination spider good than spider - bad combination than the not spider fearful control group. Despite the fact that there was no difference between these two groups in an IAT task without spiders. The IAT results by Ellwart, Eni, and Rinck (2006) showed longer RTs to spider – pleasant words than spider-unpleasant words, and this effect was somewhat stronger for the spider fearful group than the control group with individuals not fearful for spiders. Combining these three experimental paradigms a research quest set out to investigate how fear of bears and wolves affects physiological activity, performance efficiency, and associations. Thus, in the present study the participants first got a picture viewing task, without any additional task, not to risk that behavioral artifacts would obscure differential physiological responding to the different animal (Moose -Alces alces, Hare -Lepus europaeus, Wolf, and Bear) pictures. As physiological response is different for an animal one is afraid of compared to other animal pictures even when all other stimuli are animals (Flykt & Bjärtå, 2008) we choose to use only animals, and, moreover, all animals from the same geographic area. We hypothesized that exposure 9 to a feared animal would result in an increased physiological activity, and that eye movements would differ between a feared animal and not feared animals. The participants had the task to search for a hare in 3x4 pictures search arrays and decide if a hare was present or not. The other pictures in the search arrays were moose. Occasionally a bear or a wolf picture could appear somewhere in the search arrays, but the participants were instructed to disregard their presence. This is a modified version of the visual search task used by Miltner and co-workers (2004), with spider fearful participants, in which it was expected that reaction times for target in search arrays were prolonged when a feared animal oddball distractor occurred among moose (Cervus canadensis) distractors. We hypothesized that the presence of a feared oddball- distractor would prolong RTs as efficiency would be reduced. Moreover, we also hypothesized that participants would dwell longer on their feared animal, and thus explain the prolonged RTs. The participants were also presented with four short IAT tasks. Two of the tasks included bear and hare pictures and two included wolf and hare pictures, the bear-IAT and the wolf-IAT respectively. In one of the wolf-IAT and one of the bear-IAT the carnivore required the same response as for good words while in the remaining wolf- and bear-IAT required the same response as for bad words. We hypothesized that it would be harder to sort a feared carnivore with good words, than with bad words, and thus RTs would increase and accuracy would decrease in these cases. Thus, we expected that both the participants in the bear fear group and the wolf and bear fear group would have longer RTs, less accuracy, and to look more at the labels used for the sorting when bear had to be sorted with good words than when sorted with bad words (and hares with bad words), as compared to the no fear group. 10 We also expected that the wolf fearful participants (the group with participants fearing both bears and wolves) would have longer RTs, less accuracy and to look more at the labels used for the sorting when wolf pictures where sorted with good words, as compared to the other two groups (i.e. the no fear and the bear fear group). METHOD Method Participants Participants were recruited from a survey among people in areas with presence of brown bear and/or wolves in Sweden. The respondents had, amongst others, indicated if they were afraid of encountering brown bear or wolf in forest. The participants in the present study were recruited based on their responses to this survey-item and their indicated willingness to participate in a follow-up study. Thirty-nine persons accepted to participate of which 37 did eventually participate, 15 (mean age 58 years SD = 13, 5 women) not fearful of bear or wolf (NF), 8 (mean age 46 years SD = 14, 6 women) bear fearful (BF), and 14 (mean age 55 years SD = 16, 9 women) bear and wolf fearful (BWF). The groups did not differentiate on State-Trait Anxiety Inventory (STAI-T) scores (Spielberger, Gorsuch, & Lushene, 1970), F<1. Stimulus materials 11 Twelve square color pictures of each of the animal (Bear, Wolf, Moose, and Hare) were used. The pictures were mostly downloaded from different internet sites and edited to a square format. The final selection of 48 pictures was rated for how beautiful the pictures were and how harmless they were perceived. An ANOVA of a rating study with 27 participants showed no difference in how beautiful the pictures were rated, F<1, while only hare was rated as harmless (4.82), with moose as less harmless (3.66), and wolf (2.59) and bear (2.68) were rated as low on harmlessness (i.e. rated as harmful), F(3, 75) = 110.30, p < .01, mse = 0.333, 2 = .82, and Helmert contrasts; Hare vs. later F(1, 25) = 192.78, p < .01, mse = 0.457, 2 = .89, Moose vs. later, F(1, 25) = 66.83, p < .01, mse = 0.405, 2 = .73, and Wolf vs. Bear, F < 1. Apparatus The stimulus materials were presented on a 22 in. (56 cm) screen. The size of the arrays on the screen was approximately 35 cm x 20 cm, and the viewing distance was approximately 0.8 m. The SMI Experiment Center was used for the stimulus presentation, collection of eye movements was done with the iView X 250 from SMI and collection of RTs with a USB numeric keyboard as a response pad was done with the same software. The electrocardiography (ECG) signal, skin conductance and a signal indicating the presentations of the stimulus materials were recorded by the BioPac system (see, e.g., Frazier, Strauss, & Steinhauer, 2004) MP100, with the dedicated software, AcqKnowledge 3.9 (see e.g., Leong, Mann, Wallymahmed, MacFarlane, & Wilding, 2001), both provided by BioPac Systems (BioPac Systems, Inc., Goleta, CA). The ECG and skin conductance signals were sampled at 2 KHz. 12 Procedure The participants were contacted by phone. Upon arrival to the lab the participants were informed about the study, and then signed the informed consent form. Before starting the experiment a test screen simulating the experimental stimulus settings was displayed to control for the visual abilities. ECG electrodes were attached (the ground to the neck, the negative electrode on the right side about 1 dm below the armpit, and the positive electrode at about the same position on the left side) followed by SCR electrodes that were attached to the distal phalanx of the first and second finger on the non-dominant hand. To avoid artifact driven responses the participants were explicitly instructed not to press on the electrodes. . Participants were then instructed to look at the screen so that the experimenter could control that the eye tracking camera could detect the eyes (i.e. that the eyes were at the level of the camera and that the participants eyes were at the right distance. If not, the screen level and the participants distance to the screen were adjusted until the camera reliably detected the eyes. The participants were encouraged to move their heads as little as possible during the experiments. Experiment 1: Picture viewing For Experiment 1 the participants were informed that they would look at animal pictures showing bears, wolves, moose, and hares, and there would be one picture on the screen at a time preceded by a circle that should be fixated. The participants were instructed to look at the picture during the entire exposure time. First the eye tracking camera was calibrated. For this purpose the participants were asked to steadily focus their gaze on a 13 white circle that moved across the screen. The same procedure reoccurred as a validation at the end of the experiment. The exposure time was approximately 4000 ms long. Forty animal pictures equally distributed for the different animal categories were shown with an inter trial interval of 10 seconds. Experiment 2: Visual search In Experiment 2 the participants were first shown an external USB number keypad used for response collection and were demonstrated how to hold the keypad with both hands to be able to respond with their thumbs without pressing the SCR electrodes against the keypad causing artifact driven responses. Then participants were informed that they would be presented matrices with 12 animal pictures and these animals could be moose, hares, wolves, and bears. An example of a matrix (3 row x 4 columns) with only moose was shown. If a picture of hare was present in the matrix they were instructed to press the defined response button on the keypad with their dominant hand and that if no hare picture was present in the matrix they should press another predefined response button on the keypad with their non-dominant hand. This asymmetry was used because the focus of interest was on the arrays with a target picture. A second matrix example was shown with a hare picture among eleven pictures of moose. It was emphasized that it was only if there was a hare picture in the matrix that they should respond with their dominant hand, and that they should ignore pictures of wolves and bears. Matrices were exposed for two seconds and the participants should attempt to respond as fast possible without jeopardizing the accuracy. The participants were encouraged to try to respond while the 14 matrix still was present on the screen. Thereafter the eye tracking camera was calibrated as described for the picture viewing. Seventy-two matrices with an inter trial interval of 10 seconds were presented. Twelve matrices showed twelve moose pictures, 12 matrices showed eleven moose pictures and one hare picture, 12 matrices showed eleven moose picture and one wolf picture, 12 matrices showed eleven moose picture and one bear picture, 12 matrices showed eleven moose pictures, one hare picture and one wolf picture, 12 matrices showed eleven moose pictures, one hare picture and one bear picture. For the later two combinations the placement of the wolf and bear pictures in relation to the hare picture was counterbalanced. Only target trials were analyzed as the no target trials were considered filler trials. Experiment 3: IAT The participants were instructed that the third experiment would consist of four parts and that they would be informed by the experimenter before the start of each part. An example of the screen with a different animal category in each upper corner of the screen paired with the word good or bad (in Swedish) was displayed to the participants. The participants were informed that the combination of the category words (e.g. bear good) in the upper corner of the screen would be shown constantly during the experiment. Then the participants were informed that a word or an animal picture would occur in the lower centre of the screen and were shown examples. One example with an animal picture and one with a word were shown. The participants were instructed that their task was to sort the animal pictures and the words in accordance with the categories in the upper corners of the screen. Examples like: If there was a bear picture and the category bear was written 15 in the upper right corner the participant should respond with the predefined response button on the right side of the keypad. If there was a word like wonderful (in Swedish) and the category good words was displayed in the right upper corner of the screen the participant should respond with the predefined response button on the right side of the keypad. Then the participants were presented to a picture showing a list of good words and a list with bad words in two columns. The words from these two lists were the words to be used during the experiment. The information to the participants was repeated until the participants were confident about their task. The participants were also informed that before each presentation of an animal picture or a word a white circle would occur announcing the advent of the presentation. The participants were encouraged to respond as fast as possible without jeopardizing the accuracy. The eye tracking camera was calibrated, a first part was run, and validation of the gaze tracking was made. Before the next part the participants were informed that the animal categories may have changed as well as that the location of categories in the upper corners of the screen. It was however emphasized that this new setup would be the same during the entire part. The camera was calibrated again before the start of the IAT task and validation was made after. The same procedure was repeated for the remaining two parts. Each part consisted of 32 trials of which half were animal pictures (equal amount of each animal) and half were Swedish words (half good – corresponding to; nice, happy, lovely, wonderful, peaceful, excellent, pleasant, and enjoyable, and half bad words- corresponding to; failure, awful, painful, evil, mean, terrible, violent, and angry). The order of the four parts was counter balanced. Inter trial intervals were approximately 4000 ms. 16 The experimenter recommended the participants to rest their eyes between each experiment and between the parts of the IAT task. After the experiments the participants received a cinema ticket for a movie of their own choice and a certificate of value for a course on training hunting dogs to track bears or a course on training dogs to ignore wolf tracks. (Both courses were relevant for people living in the regions where the study was conducted.) Both courses were given at the Swedish wildlife damage centre. The value certificates were not personal but could be transferred to someone else. Data Treatment and Analyses The peaks of the R-waves were used for the calculation of the interbeat intervals (IBIs). All trials were visually inspected and artifacts were removed. No attempts to correct for respiratory responses were made, as the HR was used as an index of orienting response (OR) and defense response (DR), and respiration is a central aspect of the OR (Stekelenburg & van Boxtel, 2002). The baseline consisted of the two last entire IBIs before stimulus onset, and change scores were calculated for the five IBIs following the IBI containing the stimulus onset. For two participants in the not fearful group HR change score data from a few trials were deleted as the participants had double-beats and the difference from one IBI to another became unreasonable (some more than 100 bpm) these trials were omitted from further analysis as they must be seen as outliers. The omitted data came from different conditions. Only the correct responses were used for the analysis of RT data. SCRs were scored semi-manual on screen. Two different measures of SCRs were used to capture different aspects of sympathetic activity evoked by the stimulus materials. SCR magnitude was used to establish the average deflection from 17 SCL for the largest SCR elicited in the time frame 0.9 to 4 s for the condition, while number of elicited responses is the number of deflections elicited during the same time frame per condition. From eye movement data total (i.e. accumulated) length of fixation and the fixation counts for specified regions of interest were used. The time not captured by the total length of fixation was time when the eyes moved or when they were fixated outside the specified areas of interest. All dependent measures were analyzed with separate ANOVAs for the respective experiments. RESULTS Experiment 1: Picture viewing The 3 group (bear fearful, wolf and bear fearful, not fearful) x 4 animal pictures (bear, wolf, moose, hare) ANOVA for number of elicited SCRs showed that more SCRs were elicited to bear pictures than the other pictures, as shown by a main effect of animal picture, F(3, 102) = 4.61, p < .01, 2 = .12, mse = 1.315, and a Helmert contrast, F(1, 34) = 9.59, p < .01, 2 = .22, mse = 1.877. The corresponding ANOVA for SCR magnitude did not show any effect (despite the fact that the picture sequence was counter balanced across subjects). The 3 group (bear fearful, wolf and bear fearful, not fearful) x 4 animal pictures (bear, wolf, moose, hare) x 5 IBI ANOVA for heart rate change scores in BPM only showed that the BF group had a general decrease in bpm over the five heart beat, while no such decrease was shown in the other two groups, as shown by an interaction effect between group and IBI, F(8, 136) = 2.46, p < .05, 2 = .13, mse = 2.42 (see Figure 1). To control if this difference could be explained by a stronger deceleration to 18 their feared animal (i.e. bear), a follow-up analysis was conducted. An Helmert contrast on a 4 animal pictures (bear, wolf, elk, hare) x 5 IBI ANOVA for heart rate change scores for the bear fearful group showed tendencies for bear pictures to generate more deceleration (-1.18) -than wolf (-0.64), elk (-0.60), and hare (-.26) pictures, F(1, 7) = 2.28, p < .08 (directed hypothesis), 2 = .25, mse = 8.21. The 3 group (bear fearful, wolf and bear fearful, not fearful) x 4 animal pictures (bear, wolf, elk, hare) ANOVA for the fixation time showed that the participants had a tendency to fixate bear pictures (2674 ms) longer than elk pictures (2496 ms) (Main effect of animal picture, F(3, 102) = 2.73, p = .06, 2 = .07, mse = 91275, and Bonferroni corrected pairwise comparisons, p = .02). The corresponding ANOVA for fixation count showed that the participants fixated a larger number of times on bear pictures (than the wolf pictures). This was shown by a tendency to a main effect of animal picture, F(3, 102) = 2.44, p < .09, 2 = .07, mse = 90.84, and Bonferroni corrected pairwise comparisons (p = .03). These results suggest that pictures of bears are special in that gaze is more often returned to bears than other animals, the total viewing time then became longer. Bears also elicit more sympathetic activation as seen as the number of SCRs. Moreover, the result of the additional analysis of heart rate change scores suggest that bear fearful individuals are also responding with an increased parasympathetic activity to bear pictures. Experiment 2: Visual Search The 3 group (not fearful, bear fearful, wolf and bear fearful) x 3 oddball distractor type 19 (not present, bear, wolf) ANOVA for response accuracy showed that more correct answers were given for search arrays with a target without an oddball distractor (9.56), during the two seconds the participants had for their responding, than for the search arrays that contained an oddball distractor (bear = 8.60 wolf =8.35). This was shown by a main effect of oddball distractor, F(2, 64) = 8.31, p < .01, 2 = .21, mse = 1.76, and Bonferroni corrected comparisons (the differences had ps < .01). No other effects were shown for accuracy. The 3 group (not fearful, bear fearful, wolf and bear fearful) x 3 oddball distractor (not present, bear, wolf) ANOVA and subsequent a-priori contrasts for RTs showed in accordance with the hypothesis that (see Figure 2): The BWF group showed prolonged RTs to search arrays with an oddball distactor independent if it was a bear or a wolf. The BF group showed a border line effect for longer RTs to search arrays with a bear as an oddball distractor than for a search array with no oddball distractor and search arrays with a wolf as an oddball distractor. Moreover, for this group search arrays with a wolf distractor resulted in shorter RTs than for search arrays with only a target and no oddball distractor. The NF group did not show any differences in response time whether there was an oddball distractor present in a search array with a target or not. This was shown by a tendency to an interaction between group and oddball distactor, F(2, 64) = 2.42, p = .06, 2 = .13, mse = 13608, and Helmert contrasts corresponding to the hypothesis for the group (not fearful group F < 1, bear fearful group, F (1, 7) = 3.06, p = .06, 2 = .30, wolf and bear fearful group, F (1, 12) = 8.65, p = .01, 2 = .42). The 3 group (not fearful, wolf and bear fearful, bear fearful) x 3 distractor type 20 (not present, bear, wolf) of number of skin conductance response (SCRs) showed more elicited responses to search arrays with a bear as an oddball distractor than when there was no oddball distractor in the search arrays. This was shown by a main effect of distractor type F(2, 68) = 4.11, p < .03, mse = 1.948, 2 = .11, and Bonferroni corrected comparison (p < .01). The corresponding analysis for the magnitude of skin conductance responses (SCRs) showed no main effects. The 3 group (not fearful, wolf and bear fearful, bear fearful) x 3 distractor type (not present, bear, wolf) x 5 interbeat interval (IBI) after stimulus onset ANOVA of heart rate changes showed that an increase over IBIs (from -0.57 to 0.35), as shown by a main effect of IBI, F(4, 136) = 8.52, p < .01, mse = 5.95, 2 = .20. Moreover, a tendency for the BWF group to have a larger deceleration (-1.00) bpm) than for the NF group (0.60) was shown, F(2, 34) = 3.13, p < .07, mse= 48.51, 2 = .16, and Bonferroni corrected t-test p < .07. The 3 group (not fearful, wolf and bear fearful, bear fearful) x 3 distractor type (not present, bear, wolf) ANOVA of fixation count on the target showed that the BWF group showed fewer fixations (1.47) on the target than the NF group (2.04). This was shown by a main effect of group, F(2, 32) = 4.44, p = .02, 2 = .22, mse = 0.280, and Bonferroni corrected pairwise comparisons (p < .03). The targets presented without an oddball distractor (1.93) or a bear distractor (1.88) were fixated more often than targets presented with a wolf distractor (1.66), which was shown by a main effect of distractor type F(2, 64) = 6.13, p < .01, 2 = .16, mse = 0.109, and Bonferroni corrected pairwise comparisons (p < .03 and p < .06 respectively). The 3 group (not fearful, bear fearful, wolf and bear fearful) x 3 distractor type 21 (not present, bear, wolf) ANOVA of fixation time on the target showed that the target is fixated during longer time if there is no oddball distractor (448 ms) than if it is a wolf (387 ms). There was also a tendency to fixate the target less if a bear oddball distractor was present (431 ms). This was shown by a main effect of distractor type, F(2, 64) = 6.83, p < .01, 2 = .18, mse = 4905, and Bonferroni corrected pairwise comparisons (p < .01 and p<.06 respectively). The 3 group (not fearful, bear fearful, wolf and bear fearful) x 2 distractor (bear, wolf) ANOVA of fixation count on the distractors showed that wolf oddball distractors were fixated more often (0.89 times) than bear distractors (0.64 times). This was shown by a main effect of distractor, F(1, 32) = 18.08, p < .01, 2 = .36, mse = 0.055. The 3 group (not fearful, bear fearful, wolf and bear fearful) x 2 distractor (bear, wolf) ANOVA of fixation time on the distractors showed that wolf oddball distractors were fixated longer (153 ms) than bear distractors (107 ms). This was shown by a main effect of distractor, F(1, 32) = 17.93, p < .01, 2 = .36, mse = 1954. These results suggest that in general the participants spend more time looking at an oddball distractor when it is a wolf than when it is a bear, while the bear oddball distractor seems to have elicited the most sympathetic activation. An oddball distractor reduces the viewing time of the target as compared to when there is no oddball distractor. That did correspondingly reduce the viewing of the target, which lead to fewer correct responses (during the postulated time) for search arrays with an oddball distractor than without. Experiment 3: Bear-IAT 22 For the IAT tasks that presented bears and hares, a 3 group (no fear, bear fear, or wolf and bear fear) x 2 sorting of bear (good or bad words) x 2 animal picture (bear or hare) of the RTs showed that only the BWF group had slower responses when bear was sorted with good words (871 ms) than with bad words (804 ms) while this effect was not present in the not fearful (good words 817 ms bad words 830 ms) or the BF group (good words 813 ms bad words 837 ms). This was shown by an interaction effect between group and sorting of bear, F(2, 31) = 4.83, p < .02, 2 = .24, mse = 5832, and subsequent simple contrasts. The corresponding ANOVA for correct answers during the second available for responding showed that bear pictures was responded at to a larger extent (5.34 times) than was hare pictures (4.51 times), as shown by an main effect of animal picture, F(1, 31) = 6.83, p <. 02, 2 = .18, mse = 3.26. This is mainly due to that less accuracy occurs for Hare pictures when they should be sorted with bad words (3.91- 49%) than when sorted with good words (5.11 – 64%), this was shown by an interaction effect between sorting of bear and animal picture, F(1, 31) = 6.61, p < .02, 2 = .18, mse = 1.66. In a 3 group (no fear, bear fear, or wolf and bear fear) x 2 sorting of bear (good or bad words) x 2 animal picture (bear or hare) of number of elicited SCRs it was shown that sorting bear with good words and hare with bad words elicited more SCRs (1.17) than when categorizing bear with bad words and hare with good words (0.70). This was shown by a main effect of sorting of bear, F(1, 33) = 5.87, p < .03, 2 = .15, mse = 1.28. In a 3 group (no fear, bear fear, or wolf and bear fear) x 2 sorting of bear (good or bad words) x 2 animal picture (bear or hare) of SCRs magnitude it was shown that the BWF group showed smaller responding to bear 23 than hare pictures when sorted bear sorted with bad words and hare sorted with good words, while no such difference was shown for any other condition. This was shown by a three-way interaction effect between group, sorting, and animal picture, F(2, 33) = 3.83, p < .04, 2 = .19, mse < 0.001, and subsequent follow-up tests, t(13) = 2.37, p < .04, d = .0.84. The 3 group (no fear, bear fear, or wolf and bear fear) x 2 sorting of bear (good or bad words) x 2 animal picture (bear or hare) for the HR change scores between the IBI before and the IBI after the IBI with the stimulus onset showed less acceleration for bear pictures (0.56 bpm) than hare pictures (1.11 bpm) when bear was sorted with bad words and hare pictures sorted with good words, while no such difference between bear (1.30 bpm) and hare (0.93 bpm) when bear was sorted with bad words and hare was sorted with god words. This was shown by an interaction effect between sorting and animal picture, F(1,33) = 4.98, p < .04, 2 = .13, mse = 1.44, and subsequent contrasts, t(35) = 2.05, p < .05, d = 0.35, and t(35) = 1.16, n.s. In the 3 group (not fearful, bear fearful, or wolf and bear fearful) x 2 sorting categories (bad or good words) x 2 animal (bear or hare) ANOVAs of fixation count and fixation time it was shown that in the group fearful of both wolves and bears the participants had a tendency to, fixate more and longer on the area displaying the words; good words - bears when exposed to bears pictures than good words – hares when exposed to hare pictures, and on the area displaying bad words – hares when exposed to hare pictures than bad words – bears when exposed to bears. This was shown by a tendency to an interaction effect between group, category, and animal 24 for number of fixations of the words supporting the sorting, F(2, 33) = 3.00, p < .07, 2 = .15, mse = 0.184, and for the fixation time, F(2, 33) = 2.79, p < .08, 2 = .15, mse = 6404, and a-priori simple contrasts showing tendencies or differences for the group fearful of both wolves and bears, while no such tendencies or differences where show for the other two groups. These results show that when the sorting does not lead to cognitive dissonance smaller and fewer responses are elicited than when the sorting leads to a cognitive dissonance. This effect appears in different measures and sometimes as a general effect and sometimes as a group specific effect. Experiment 3: Wolf-IAT For the IAT tasks that presented wolves and hares, a 3 group (no fear, bear fear, wolf and bear fear) x 2 sorting of wolf (good or bad words) x 2 animal picture (wolf or hare) of the RTs only tended to slower responses when wolf pictures were sorted with good words (862 ms) as compared to whenthe wolf pictures were sorted with bad words (831 ms), as shown by F(1, 30) = 3.28, p = .08, 2 = .10, mse = 9108.91. No other effects or tendencies were shown in this analysis. The corresponding ANOVA for correct answers during the second available for responding complemented the previous ANOVA by showing that less correct responses for the condition where wolf should be sorted with good words and hare with bad words, as shown by a main effect of sorting of wolf pictures, F(1, 30) = 9.81, p < .01, 2 = .25, mse = 6.75. No other effects or tendencies were shown. 25 In the 3 group (no fear, bear fear, or wolf and bear fear) x 2 sorting of bear (good or bad words) x 2 animal picture (wolf or hare) ANOVAs of number of elicited SCRs no effects were shown. Neither was any effect shown by the corresponding ANOVA for SCR magnitude. The 3 group (no fear, bear fear, or wolf and bear fear) x 2 sorting of bear (good or bad words) x 2 animal picture (wolf or hare) for the HR change scores between the IBI before and the IBI after the IBI with the stimulus onset showed that the bear fearful group elicited less acceleration to trials with wolf (0.45) than hare (1.47) pictures, while the BWF group tended to elicit more acceleration to wolf (0.85) than hare pictures (0.57). No such differential responding was shown in the NF group (1.07 vs 0.95). This was shown by a interaction effect between group and animal, F(1,33) = 4.98, p < .04, 2 = .13, mse = 1.44, and subsequent contrasts, t(7) = 2.79, p < .03, d = 1.03, t(13) = -1.83, p = .09, d = 0.26, and t(13) = -0.50, n.s. respectively. In the 3 group (not fearful, bear fearful, or wolf and bear fearful) x 2 categories (bad or good words) x 2 animal (wolf or hare) ANOVA of fixation count it was shown that in the BWF group the participants had a tendency to fixate fewer times on the area displaying the words; bad words - wolves when exposed to wolves pictures than bad words – hares when exposed to hare pictures. Also the BF group tended to do this. The latter was shown by a tendency to an interaction effect between group, category, and animal for number of fixations of the words supporting the sorting, F(2, 33) = 2.61, p < .09, 2 = .14, mse = 10.93, and a-priori 26 simple contrasts showing a differences for the BWF group (p< .02), and a borderline effect for the BF group (p<.06). In the 3 group (bear fearful, not fearful, wolf or bear fearful) x 2 categories (bad or good words) x 2 animal (wolf or hare) ANOVA of fixation time it was shown a general effect of shorter average time fixating the area displaying the words; bad words – wolves when exposed to wolf pictures than any other word area included in the analysis. This was shown by an interaction effect of category and animal, F(1, 33) = 8.46, p < .01, 2 = .20, mse = 3332.03. The interaction effect also explains the main effect of category F(1, 33) = 12.62, p < .01, 2 = .28, mse = 3332.03, and the tendency of animal F(1, 33) = 3.82, p < .06, 2 = .10, mse = 1892.69. No other effects or tendencies were shown. Overall also these results suggest that congruence results in less responding, while dissonance results in more, like for the Bear IAT. Wolves seems however to be a less potent stimuli for eliciting these responses than bears are. Discussion Due to the complexity of the results (see Table 1) the discussion is divided into four parts. Each of the three first parts restates and discusses the main results from each paradigm. The last part is a general discussion, that integrates results from the different paradigms and addresses them in relation to fear of snakes and spiders. Moreover, the indentified differences between fear of bear and fear of wolf is discussed. Discussion of Picture viewing results 27 In the picture viewing most SCRs, most eye fixations, and the longest eye fixations were shown for bear pictures. Moreover, the BF group showed a deceleration to bear pictures. Bear pictures seems to be potent stimuli for eliciting physiological responding independently of whether or not the participants were selected to be fearful or not fearful of bears. The HR deceleration shown to bear pictures by the BF group (but not by the BWF group) can be interpreted in two different ways. One is that, in areas where brown bear occur (mainly the northern hemisphere), fear for bears is a more rational fear than fear for snakes and spiders. Fear for snakes or spiders elicit more acceleration of heart rate to pictures of the feared animal as compared to not feared animals (see e.g. Flykt & Bjärtå, 2008). The larger deceleration of heart rate to bears in the BF group is contrary to results from snakes and spider fearful persons (see e.g. Flykt & Bjärtå, 2008), but may be indicative of deployment of attentional resources (Graham, 1992) directed towards the threat (i.e. the bear). This could be related to a need to keep track of movement directions of the bear. In this situation decelerating heart rate response may even be functional by to avoid a bear attack (Herrero, 2006). Another explanation would be that in the corresponding research with snake and spider fearful participants the imminence of the threatening animal is supposedly high with an enlarged spider or snake picture display. In the present study the size of the bear corresponded to seeing a bear at a distance. Thus, the difference in direction of HR change (deceleration for bears in bear fearful and acceleration for e.g. spiders in spider fearful) could in the light of Fanselow and Lester’s (1988) etiological model of defense behaviors and perceived distance to predator be due to a difference in predator (threat) imminence. 28 Accordingly the bear pictures triggering a post-encounter defensive behavior response (i.e. freezing) related physiological activity (HR deceleration), while the spider and snake pictures triggering a circa-strike defensive behavior response (escape) related physiological activity (HR acceleration). Post-encounter defensive behavior is related to HR deceleration and circa-strike defensive behavior is related to HR acceleration is supported by a proposed model for human physiological responding in relation to threat imminence from the Lang research group (Lang, Bradley, & Cuthbert, 1997). Discussion of Visual search results In the visual search task fewer correct responses were shown for search arrays containing an oddball distractor, and there were more and longer fixation on the wolf distractors. Most SCRs were elicited to search arrays with bear oddball distractors. Longer RTs were shown to search arrays with a feared distractor. Moreover, the BF group also showed faster responses to the target (i.e. a hare picture) for search arrays with a wolf oddball distractor than for targets in search arrays without any oddball distractor. The BWF group showed fewer fixations on the target. The only effect that showed a clear association between animal fear and the feared animal in the present study was the prolonged RTs to the hare targets when a feared animal was present in the search arrays (see Miltner et al., 2004). Less fixations on the targets was only shown in the BWF group. However, no fear specific increase of fixation on the feared animal was shown. The RTs effect was a replication and extension of the finding with prolonged RTs to a spider oddball distractor for spider fearful participants (see Miltner et al., 2004). Furthermore, 29 shorter RTs to targets were shown for the BF group when a wolf oddball distractor was present in the search array. This somewhat surprising result might shed some new light on the processing that occurs in visual search tasks with feared animals (see e.g. Öhman et al., 2001; Miltner et al., 2001; Flykt & Caldara, 2006). It seems like attention not only dwell on the feared oddball distractor, but that an odd ball distractor not consisting of the feared animal could enhance the processing of the target present response. One reasonable explanation is that a wolf oddball distractor serves as a safety signal for the bear fearful group. A parallel is the control stimulus (CS-) in differential classical conditioning that never follows by the unconditioned stimulus (UCS) to ensure that the increase in responding to the conditioned stimulus (CS+) is due to an associative learning effect and not a sensitization effect. The magnitude of responding to CS- tends to decrease over time as the participants understand that the CS- signals that no UCS will follow (see e.g. Lovibond, Siddle, & Bond, 1993, Figure 1, lower left panel, the conditioning phase). We argue that the bear fearful participants will be more efficient in their response as they know that they will not suddenly see their feared animal. In other words no worries of encountering the feared animal are imposing on the available resources, so that those could be more efficiently used. This result has implication for the attention control theory (Eysenck et al., 2007). which suggests that cognitive inhibition of task irrelevant information is decreased when such information is a fear eliciting stimulus. This would be due to that the feared stimulus drives a bottom-up process suppressing cognitive inhibition. The theory also suggests that shifting of resources between tasks is harder when exposed to a fear elicitin 30 stimulus. Our finding with faster RTs to a target when a not feared oddball animal distractor is present adds to the attention control theory (Eysenck et al., 2007). That is, our results suggest that the mere possibility that a fear eliciting stimulus could be present decrease the cognitive inhibition. This decrease in cognitive inhibition is not caused by the actual threat but caused by the vigilance induced by the potential appearance of this threat. In other words, our finding suggests that; the mere knowledge that a threat could appear decreases cognitive inhibition. That RTs are prolonged to target with a feared oddball distractor present in the display but that the measures of overt attention (i.e. eye movements) did not show a corresponding effect further support cognitive processing. That is, previous results from visual search with spiders and snakes as feared animals have been interpreted as faster detection (Öhman et al., 2001), longer dwell times (Miltner et al., 2004), and motor preparations (Flykt & Caldara, 2006; Flykt, 2006) cannot be the full explanation for visual search with wolves and bears. Instead we suggest that fear for large carnivores, to a large extent, is due to a cognitive process that is rather elaborated. That fear for bears and wolves would be driven by a specific fear module, as has been suggested for fear of snakes and spiders (Öhman & Mineka, 2001), is not supported by our visual search data. Discussion of IAT results In the bear-IAT the number of correct responses decreased when hare pictures had to be sorted with bad words. Furthermore, more SCRs were elicited when bear pictures had to be sorted with good words. When bear pictures had to be sorted 31 with bad words less HR acceleration was shown than when hare pictures had to be sorted with good words. Moreover, the BWF group showed longer RTs to sort bears with good words, larger SCR magnitude when bear pictures were sorted with good words, and showed more and longer fixations to the bear/good and hare/bad word pair displays than to bear/bad and hare good word pairs displays. These results are in line with the IAT-results for snake and spider fearful participants that had to sort pictures of their feared animal with good words resulted in longer RTs (Teachman, Gregg, & Woody, 2001; Teachman, & Woody, 2003). However, these effects indicating cognitive dissonance are not shown in the BF group. In the wolf-IAT all groups showed longer RTs, fewer correct responses for wolf-good and hare-bad than for wolf-bad and hare good trials and longer fixations to the earlier word pairs. This general result indicated a general negative association to wolves independently of being fearful of wolves or not. Similar results were found with spiders (Ellwart, Eni, & Rinck, 2006). These general effects are related to cognitive dissonance. As the IAT supposedly is related to attitudes (Greenwald et al., 1998) based on associations it might suggest that the BWF group has a general strong negative association to large carnivores that makes the sorting of these animals even harder than for the other two groups. They do for example have to confirm the sorting by looking at the cognitive dissonant word pairs used for the sorting more frequent and sometimes also longer. Moreover, this was also associated with larger SCRs in the bear-IAT. General discussion 32 The present results bring new and important insights for the human dimension of wildlife management. In comparison with snake and spider fear, bear and wolf fear seems to differ extensively. First and most important, in the present study the HR changes and SCRs do not reliably discriminate between the feared animal/s and the not feared animal as would been expected from studies with snakes and spiders (e.g. Öhman & Soares, 1993, 1994; Flykt & Bjärtå, 2008). The only differential HR responding was for bear pictures versus other animal pictures in the BF group. No such were shown for the BWF group, suggesting that fear for bears has different qualities in the two different groups, and thereby different etiologies. It might be that BWF group could be compared with the spider fearful (Flykt & Bjärtå, 2006) and the snake or spider fearful participants (Öhman & Soares, 1993, 1994) in previous studies as a spider fearful participant was selected based on the low fear of snake and a snake fearful participant selected based on low fear of spiders. We know of no study where the participants have been selected to be fearful of both snakes and spiders. When looking at the visual search results these neatly show a replication of the Miltner, et al. (2004) results, and the present results do not suggest that fear responding is different due to the selection of participants fearful of both bear and wolf. The group selected to be fearful of both bears and wolves showed responses in the bear - IAT that would be expected from spider or snake fearful participants exposed to spider an snakes, respectively, in IAT (Teachman, Gregg, & Woody, 2001; Teachman, & Woody, 2003). However, the group selected to be only fearful of bears did not show such effects in the bear-IAT. Thus, suggesting 33 that the fear for bears may have different etiologies if only fearful of bears or fearful of large carnivores in general. The results from the present study indicate a general tendency for bear pictures per se to result in more physiological responses as well as more and longer fixations when the only task participants was to look at the pictures. Similar general results for eye movements have also been shown for spiders (Gerdes, Pauli, & Alpers, 2009). Thus we suggest that bear pictures are potent stimuli that elicit physiological responding and more overt attention demands in a large majority of people. This activation would then be largely unrelated to if the individual report being fearful or not. It is reasonable to think that among participants in the BF group only the physiological responding elicited by bears is associated with an experience of fear. In other words, the physiological responding is appraised as part of a fear reaction in those fearful of bears only, while others do not make the similar appraisal. For the BWF group fear of bears seems to have a somewhat different ground. Instead of appraisal of physiological responses to bears it seems that in this group negative associations are the core of fear. The dissonance between good words and bears is larger for participants in the BWF group, than for participants in the BF group and the NF group. The cognitive dissonance in those fearful of both bears and wolfs is appraised as fear. Thus, fear of bears and wolves seem to be driven by elaborate cognitive processing rather than by evolutionary old dedicated systems for fear responding. Wolves are associated with negative words among a majority of people, wolf pictures do not elicit more physiological responding in tthe BWF group, than among 34 the other participants. Therefore, it seems reasonable to suggest that in those fearful of bears and wolves the negative associations have been appraised differently than for those not fearful of wolves. It also indicates that the etiology of fear of bears and of wolves do not differ much in those fearful of both animals, but that fear for wolves and fear for bears in those only fearful of bears differs despite that both are suggested to have a cognitive origin. Such explanation is supported from different fits to a path model of fear for bears and wolfs on questionnaire data done by this research group (Johansson, Karlsson, Pedersen, & Flykt, 2012). One limitation of the present results is the low number of participants in the BF group. The small sample size might results in effects that, as a matter of fact, are present in this population do not show. That is, an effect obtained for the BWF but not for the BF group might be due to a lack of power. However, the difference in physiological responding between the groups was based on the presence of an effect in the BF group, but not in the BWF group. It would be hard to argue this is a result of the small sample size. Despite that some differences might be due to type II errors in the group only fearing bears, the differences shown based on an effect in the group only fearing bears but not in the group fearful of both animals strongly suggest a difference between these two groups. It is, indeed, important to emphasize that the etiology of fear for bear may differ depending on if the individual only fear bears or other large carnivores as well, as this can have direct consequences for wildlife management. Future research has to address this question with great care as one specific intervention might give opposite results in two groups with different etiologies. 35 Fear of bears and wolves seem to be driven by elaborate cognitive processing rather than by evolutionary old dedicated systems for fear responding. 36 References Derakshan, N. & Koster, E. H. W.(2010). Processing efficiency in anxiety: Evidence from eye-movements during visual search. Behaviour Research and Therapy, 48, 1180-1185. Ellwart, T., Rinck, M., & Becker, E. S. (2006).From Fear to Love: Individual Differences in Implicit Spider Associations. Emotion, 6, 18-27. Ericsson, G., & Heberlein, T. A. (2003). Attitudes of hunters, locals, and the general public in Sweden now that the wolves are back. Biological Conservation, 111, 149-159. Ericsson, G., Sandström, C., Kindberg, J., & Stoen, O-G. (2010). Om svenskars rädsla för stora rovdjur, älg och vildsvin. [On Swedes’ fear of large carnivores, moose and wildboar]. Report 2010:1. Umeå: Swedish University of Agricultural Sciences. Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and cognitive performance: Attentional control theory. Emotion, 7, 336-353. Fanselow, M. S. & Lester, L. S. (1988). A functional behavioristic approach to aversively motivated behavior: Predatory imminence as a determinant of the topography of defensive behavior. In R. C. Bolles & M. D. Beecher (Eds.), Evolution and learning, (pp. 185-212). Hillsdale, N J: LEA. Flykt, A. & Bjärtå, A. (2008). The time course of resource allocation in spider-fearful participants during fear reactions. Cognition & Emotion, 22, 1381-1400. Flykt, A. (2006). Preparedness for action: Responding to the snake in the grass. American Journal of Psychology, 119, 29-44. 37 Flykt, A., & Caldara, R. (2006). Tracking fear snake and spider fearful participants during visual search: A multiresponse domain study. Cognition and Emotion, 20, 1075-1091. Frazier, T. W., Strauss, M. E., & Steinhauer, S. R. (2004). Respiratory sinus arrhythmia as an index of emotional response in young adults. Psychophysiology, 41, 75–83. Fredrikson, A., M., Annas, P., Fischer, H. & Wik, G. (1996). Gender and age differences in the prevalence of specific fears and phobias. Behaviour Research and Therapy, 34, 33-39 Gerdes, A. B. M., Pauli, P., & Alpers, G. W. (2009). Toward and away from spiders: Eye-movements in spider-fearful participants. Journal of Neural Transmission, 116, 725-733. Gerdes, A.B.M., Alpers, G. W., & Pauli, P. (2008). When spiders appear suddenly: Spider-phobic patients are distracted by task-irrelevant spiders. Behaviour Research and Therapy, 46, 174-187. Graham, F. (1992). Attention: The heartbeat, the blink, and the brain. In B. A. Campbell, H. Hayne, R. Richardson, & B. A. Campbell (Eds.), Attention and information processing in infants and adults. Perspectives from human and animal research, 3-29. Hillsdale, NJ: Erlbaum; 1992. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74, 1464-1480. 38 Isbell, L.A. (2006). Snakes as agents of evolutionary change in primate brains. Journal of Human Evolution, 51, 1-35. Johansson, M., Pedersen, E., Karlsson, J., & Flykt, A. (submitted) Factors governing human fear of brown bear and wolf. Manuscript submitted Johansson, M., Sjöström, M., Karlsson, J., & Brännlund, R. (in press). Is human worry affecting public support of management and conservation of large carnivores? Society & Natural Resources. Karlsson, J., Bjärvall, A., & Lundvall, A. (1999). Svenskarnas inställning till varg – en intervjuundersökning [Swedes’ attitudes towards wolf – an interview study]. Report 4933. Stockholm: The Swedish Environmental Protection Agency. Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1997). Motivated attention: Affect, activation, and action. In P.J. Lang, R. F.Simons, & M. T. Balaban (Eds.), Attention and orienting: Sensory and motivational processes. (pp. 97-135). Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers. Leong, K. S., Mann, P., Wallymahmed, M., MacFarlane, I. A., & Wilding, J. P. H. (2000). Abnormal heart rate variability in adults with growth hormone deficiency. Journal of Clinical Endocrinology & Metabolism, 85, 628–633. Linnell J. D. C., Solberg, E. J., Brainerd, S., Liberg, O., Sand, H., Wabakken P., & Kojola, I. (2003). Is the fear of wolves justified? A Fennoscandian perspective. Acta Zoologica Lituanica: 13, 27-33. Lipp, O. V. & Waters, A. M. (2007). When danger lurks in the background: Attentional capture by animal fear-relevant distractors is specific and selectively enhanced by animal fear. Emotion, 7, 192-200. 39 Lovibond, P. F., Siddle, D. A., & Bond, N. W. (1993). Resistance to extinction of fearrelevant stimuli: Preparedness or selective sensitization? Journal of Experimental Psychology: General, 122, 449-461. Miltner, W. H. R., Krieschel, S., Hecht, H., Trippe, R., & Weiss, T. (2004). Eye movements and behavioral responses to threatening and nonthreatening stimuli during visual search in phobic and nonphobic subjects. Emotion, 4, 323-339. Öhman, A., & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an evolved module of fear and fear learning. Psychological Review, 108, 483-522. Öhman, A., & Soares, J. J. F. (1993). On the automatic nature of phobic fear: Conditioned electrodermal responses to masked fear-relevant stimuli. Journal of Abnormal Psychology, 102, 121-132. Öhman, A., & Soares, J. J. F. (1994). "Unconscious anxiety"; Phobic responses to masked stimuli. Journal of abnormal psychology, 103, 2, 231-240. Öhman, A., Flykt, A., & Esteves, F. (2001). Emotion drives attention: Detecting the snake in the grass. Journal of Experimental Psychology: General, 130, 466–478. Sabatinelli, D., Bradley, M. M., & Lang, P. J. (2001). Affective startle modulation in anticipation and perception. Psychophysiology, 38, 719-722. Seligman, M. E. (1970). On the generality of the laws of learning. Psychological Review, 77, 406-418. Seligman, M. E. (1971). Phobias and preparedness. Behavior Therapy, 2,307320. 40 Spielberg, C. D., Gorsuch, R., & Lushene, R. (1970). Manual for the State-trait anxiety inventory. Palo Alto, Calif.: Consulting Psychologists Press. Stekelenburg, J. J., & van Boxtel, A. (2002). Pericranial muscular, respiratory, and heart rate components of the orienting response. Psychophysiology, 39, 707-722. Teachman, B. A., & Woody, S. R. (2003). Automatic processing in spider phobia: Implicit fear associations over the course of treatment. Journal of Abnormal Psychology, 112, 100-109. Teachman, B., Gregg, A. P., & Woody, S. R. (2001). Implicit associations for fearrelevant stimuli among individuals with snake and spider fears. Journal of Abnormal Psychology, 110, 226-235. 41