Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S1. Overall distribution of whisk cycle phases of touch events for all units with rapid
response to touch. Plot of the relative frequency of touch events at different phases of the whisk
cycle. The blue horizontal line indicates the mean probability of touch over all phase intervals.
The shaded region around this line indicates 95 % confidence limits for a uniform distribution.
Although touch events near retraction are significantly less frequent, a sufficient number of
these events occur in each session to produce stable touch response estimates, e.g., the mean
number of touch events per session for the interval at the start of protraction (the lowest bar on
the plot) is 12 touches per session. In general, each interval had at least 8 touches per
individual session in the data sets that we analyzed.
Nature Neuroscience: doi:10.1038/nn.2283
Probability of Touch
0.2
Protraction
Protracted
Retraction
0.1
0
−π
Nature Neuroscience: doi:10.1038/nn.2283
0
π/2
−π/2
Phase in Whisk Cycle (radians)
π
Figure S1 - Curtis and Kleinfeld
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S2. Example of raw data for a unit with a touch response modulated by whisking. Plots
of high speed video frames, videographic position signal, rectified touch sensor signal, rectified
(and low pass filtered) EMG signal, and S1 cortex recording channels surrounding two touch
events during a recording session. Dots indicate times of contact events and sorted spike times
for two units (waveforms and autocorrelations are shown at the bottom). The left plots show the
rat touching the sensor during retraction and touch in the right plots is during protraction. Unit 1
is rapidly excited by touch and Unit 2 does not respond to touch. The spike activity of Unit 1 is
modulated by free whisking and its touch response is modulated by whisking. The preferred
phases of free whisking and touch for Unit 1 are near retraction, i.e., φwhsik = 0.84 π for whisking
and φtouch = 0.82 π for touch. The touch event to the left is within the preferred phase of Unit 1 as
can be seen by increased spiking briefly after touch is initiated. The touch event to the right
occurs nearly opposite to the preferred phase of Unit 1 and spiking briefly after touch remains
similar to background activity.
Nature Neuroscience: doi:10.1038/nn.2283
Epoch with Contact Near Retraction
θ = 161
o
θ = 160
o
θ = 144
θ = 134
o
Epoch with Contact Near Protraction
θ = 137
o
o
θ = 141o
θ = 154o
θ = 159o
θ = 150o
θ = 137o
Video frames
Video frames
Touch
160
o
160o
Angle, θ
130o
130o
EMG
100 µV
S1 wire 2
S1 wire 1
Multi-units
Unit 2
Unit 1
50 ms
Unit 2
Unit 1
Spikes / Bin
Stereo
waveform
100 µV
500 µs
50
0
Auto20
correlation
-20
0
20
Time from spike (ms)
0
-20
0
20
Time from spike (ms)
Nature Neuroscience: doi:10.1038/nn.2283
Figure S2 - Curtis and Kleinfeld
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S3. Distribution of touch response widths. Histogram of the width of the response to
touch (full width at half maximum amplitude) for all excited or inhibited touch responses, i.e.,
rapid excited, slow excited, and slow inhibited. Units with no response to touch are not included.
Note that all rapid excited touch response widths fall within the first bin of the plot.
Nature Neuroscience: doi:10.1038/nn.2283
40
Number of Single Units
35
Rapidly
Excited
30
25
20
Slowly Excited
and Inhibited
15
10
5
0
0
0.2
0.4
0.6
1.0
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Touch Response Width at Half Maximum Amplitude (s)
Nature Neuroscience: doi:10.1038/nn.2283
Figure S3 - Curtis and Kleinfeld
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S4. Free whisking modulation depths and spike waveform widths for all units according
to touch response type. (Top) The distribution of different classes of single units according to
the width, measured from peak to trough, of the waveform for the unit. The inserts show typical
waveforms. (Lower) Free whisking modulation depth, measured from minimum to maximum
divided by mean spike rates, plotted as functions of spike waveform width for the three
difference responses, i.e., rapid excited, slow excited, and slow inhibited, as well as
unresponsive cells.
Nature Neuroscience: doi:10.1038/nn.2283
0
Rapidly
Excited
1
Spike Rate
Count
20
Mean
Max
Min
0
-π
π
0
Phase (radians)
Free Whisk Modulation Depth
0
Slowly
Inhibited
1
0
Slowly
Excited
1
0
No Touch
Response
1
0
Nature Neuroscience: doi:10.1038/nn.2283
0
500
1000
Spike Waveform Width (μs)
Figure S4 - Curtis and Kleinfeld
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S5. Probability of burst generation by single units in S1 vibrissa cortex. To quantify
bursting activity in neurons excited by touch as opposed to free whisking in air, we produced
separate interspike interval (ISI) histograms (bin width < 0.5 ms) for spikes that occurred during
touch versus free whisking trials. To model normal spike activity, we used a generic form of the
Inverse Gaussian equation to capture the refractory period spiking and exponential decay in
number of ISI’s at higher values. Bursting activity was modeled as a Gaussian function where
the center and width were chosen manually such that the function was confined to the relative
refractory period. If a particular ISI histogram contained a distinct peak at small values, which is
indicative of bursting, the center and width of the Gaussian function were adjusted to best fit the
peak. Functions were fit to the ISI histograms using Poisson distributed maximum likelihood
estimates. Since particular ISI distributions were modeled as a linear combination of the
Gaussian and Inverse Gaussian functions, the number of ISI values contributed by each could
be calculated separately. Thus the number of ISI values that resulted from bursting was
calculated by taking the difference in number of ISI values between the Gaussian and Inverse
Gaussian fitted curves (bursting and normal spiking) and the Inverse Gaussian fitted curve
(normal spiking only). The number of ISI values due to bursting was then divided by the total
number of ISI values in order to compare the relative amount of bursting among different ISI
distributions, i.e., between free whisking and touch trials of the same neuron as well as among
different neurons.
Nature Neuroscience: doi:10.1038/nn.2283
0.12
0.08
0.04
0
0
Burst
Spikes
Count per Bin
Fraction of Spikes that Arrive
in Bursts During Touch
0.16
40
Poisson
Spikes
20
0
0
40
20
Interspike
Interval (ms)
0.16
0.04
0.08
0.12
Fraction of Spikes that Arrive
in Bursts During Free Whisking
Nature Neuroscience: doi:10.1038/nn.2283
Figure S5 - Curtis and Kleinfeld
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S6. Laminar distribution and spike widths of different classes of single units. (A)
Representative local field potential (LFP) measurements along with the current source density
(CSD) calculated from these measurements. The CSD, shown also in false color to emphasize
sinks and sources of current, was used to classify the cortical layer and depth for each
recording. The plot on the right confirms that at least the short-time measurements concur with
the past results of Swadlow, Gusev and Bezdudnaya (Activation of a cortical column by a
thalamocortical impulse. 2002. J Neurosci 22, 7766-7779). (B) Compendium of responses
across all single-units. (C) The laminar distribution of single-units that responded to active
touch, independent of their whisking related response.
Nature Neuroscience: doi:10.1038/nn.2283
A
LFP
CSD
CSD
(Swadlow et al)
Source
1
600
0
800
1000
1200
Charge Density (|e| / nm3)
400
-1
1400
1800
-100 0 100
-100 0 100
Sink
500µV
1600
10 mV / mm2
Electrode Depth (µm)
0
200
-2
-100
0 100
Time from Vibrissa Stimulus Onset (ms)
Whisking
Touch
Response Response
Yes
No
Rapidly
20 %
3%
Excited
Slowly
Inhibited
9%
4%
Slowly
Excited
15 %
13 %
None
19 %
17 %
63 %
37 %
Nature Neuroscience: doi:10.1038/nn.2283
C
Fraction of Units
B
1.0
Supragranular
Granular
0
20
40
Time (ms)
Infragranular
0.5
0
400
600
800
1000 1200 1400 1600
Electrode Depth (µm)
Figure S6 - Curtis and Kleinfeld
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S7. Examples of whisking activity that surrounds touch events. Each plot shows
whisking activity that surrounds touch events that are sorted by where touch occurs in the whisk
cycle. The data applies to the three units included in figure 4. Single whisking trajectories are
thin lines overlaid in grey and mean trajectories are thick black lines. The lower plots show the
relative frequencies of touch at different whisk cycle phases for each unit’s recording session.
Nature Neuroscience: doi:10.1038/nn.2283
EMG Activity Surrounding Touch Events
at Different Phases of the Whisk Cycle
Unit 1
-π
Unit 3
Unit 2
−3π/4
Phase in Whisk Cycle (radians)
−π/2
−π/4
0
π/4
π/2
3π/4
-100
0
100
-100
0
100
-100
Time from Touch (ms)
0
100
Number of
Touch Events
Distributions of Whisk Cycle Phases for All Touch Events
20
10
0
-π
-π/2
0
π/2 π
20
40
10
20
0
-π
-π/2
0
0
π/2 π -π -π/2
0
π/2 π
Phase in Whisk Cycle (radians)
Nature Neuroscience: doi:10.1038/nn.2283
Figure S7 - Curtis and Kleinfeld
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S8. Free whisking and touch responses for all units with rapid response to touch that are
significantly modulated by phase in the whisk cycle. Autocorrelations and waveforms for each
neuron are plotted at the right end of each row. (A1–A4). Average spike response histograms in
response to free whisking centered on ∇EMG peaks with sinusoid fits (green lines) and
resultant parameters, i.e., DC = average spike rate across histograms, Amp = amplitude of
sinusoid fit (DC to peak), φ = preferred phase, mod = response modulation depth. (B1–B4)
Average spike responses to touch with BARS fits (green lines). Peak and baseline rates
calculated from smooth curve are listed above each plot. (C1–C4) Left: Average touch response
histograms parsed by phase in the whisk cycle using ∇EMG data to determine the phase. Right:
Plots of the peak values of the touch response from fits to each of the eight intervals of the
smooth curve touch responses in the left hand panels. The parameters listed above each plot
were estimated by fitting a sinusoid to each curve and have the same designation as in panel A.
Dotted lines indicate 95 % confidence intervals for mean spike rate estimates. (D1–D4) Left:
Average touch response histograms parsed by phase in the whisk cycle using videographic
data to determine the phase. Right: Plots of the peak values of the touch response from fits to
each of the eight intervals of the smooth curve touch responses in the left hand panels. (E1–E4)
Left: Average touch response histograms parsed by speed in the whisk cycle using
videographic data to determine the speed. Right: Plots of the peak values of the touch response
from GLM fits to each of the eight intervals of the smoothed touch responses in the left hand
panels. (F1–F4) Left: Average touch response histograms parsed by angular position in the
whisk cycle using videographic data to determine the position. Right: Plots of the peak values of
the touch response from GLM fits to each of the eight intervals of the smoothed touch
responses in the left hand panels.
Nature Neuroscience: doi:10.1038/nn.2283
Spike Rate (Hz)
B1 Trial Averaged Touch
Response
Peak = 48, Baseline = 12
-π
0
π
100
100
100
100
DC = 85, Amp = 48, f = 0.41π, Mod = 1.1
0
100
DC = 38, Amp = 22, f = 0.54π, Mod = 1.2
0
DC = 54, Amp = 31, f = 0.30π, Mod = 1.0
0
DC = 73, Amp = 32, f = 0.97π, Mod = 0.88
0
DC = 54, Amp = 35, f = −0.87π, Mod = 1.3
0
50
C1 Touch Response versus phase in D1 Touch Response Versus Phase in
Whisk cycle (from EMG)
Whisk Cycle (from video)
50 100
DC = 51, Amp = 30, f = −0.76π, Mod = 1.2
0
100
100
DC = 75, Amp = 27, f = 0.98π, Mod = 0.74
0
100
DC = 56, Amp = 37, f = 0.27π, Mod = 1.3
0
DC = 40, Amp = 24, f = 0.44π, Mod = 1.2
0
DC = 83, Amp = 48, f = 0.36π, Mod = 1.2
0
100
DC = 69, Amp = 29, f = 0.91π, Mod = 0.84
0
100
100
DC = 68, Amp = 26, f = 0.82π, Mod = 0.76
0
Velocity of Vibrissa Motion (deg / ms)
E1 Touch Response Versus Velocity of
Whisking Motion
DC = 56, P 2T = 33, P kVel = 0.53, Mod = 0.59
0
50 100
DC = 52, P2T = 30 Ang = 90, Mod* = 0.57
0 50
100
100
F1Touch Response Versus Vibrissal
Position
80
100
1
0
DC = 72, P2T = 12, Ang = 90, Mod* = 0.33
0
DC = 60, P2T = 21, Ang = 125, Mod* = 0.35
0
100
DC = 40, P2T = 18, Ang = 142, Mod* = 0.45
0
100
DC = 85, P2T = 51, Ang = 156, Mod* = 0.60
0
100
DC = 68, P2T = 11, Ang = 126, Mod* = 0.18
0
100
Spike Rate (Hz)
DC = 81, P2T = 43, Ang = 138, Mod* = 0.53
0
100
-100
0
100
Time from Contact (ms)
20
0
0
50
20
0
50
0
20
0
50
0
0
10
500 µs
−20 0 20
Time from Spikes (ms)
Figure S8 (1-7) - Curtis and Kleinfeld
140
130
120
160
120
160
140
120
140
120
100
140
120
100
120
100
80
120
DC = 74, P 2T = 34, P kVel = -0.38, Mod* = 0.46
0
100
Spike Rate (Hz)
DC = 81, P2T = 31, Vel = -0.20, Mod* = 0.38
0
100
DC = 66, P 2T = 31, P kVel = -0.98, Mod* = 0.47
0
100
DC = 82, P 2T = 63, P kVel = -1.1, Mod = 0.77
0
100
DC = 38, P 2T = 29, P kVel = -0.48, Mod = 0.76
0
100
DC = 56, P 2T = 20, P kVel = -0.98, Mod* = 0.36
0
100
-1
0
0.5
-0.5
0.5
0
-0.5
0.5
0
-0.5
1
0
-1
1
0
-1
−0.5
0
0.5
-100
0
100
Time from Contact (ms)
100 µV
A1 Trial Averaged Free
Whisking Response
50
0
Peak = 72, Baseline = 17
-π
0
π
-π
0
π
-π
0
π
-π
0
π
-π
0
π
Spike Rate (Hz)
DC = 77, Amp = 37, f = 0.32π, Mod = 0.96
0
100
-100
0
100
Time from Contact (ms)
DC = 82, Amp = 30, f = 0.38π, Mod = 0.73
100
0
Spike Rate (Hz)
Spikes / bin
DC = 11, Amp = 2.2, f = −0.85π, Mod = 0.4
10
0
DC = 16, Amp = 2.8, f = 0.85π, Mod = 0.35
50
0
Peak = 61, Baseline = 14
Peak = 40, Baseline = 16
Peak = 79, Baseline = 9.2
Peak = 69, Baseline = 18
Peak = 78, Baseline = 6
-π
0
π
-100
0
100
Time from Contact (ms)
Angle Relative to Head Centerline (deg)
20
0
0
40
DC = 12, Amp = 1.1, f = 0.35π, Mod = 0.2
10
0
0
40
DC = 12, Amp = 1.0, f = 0.48π, Mod = 0.2
10
0
0
50
DC = 9.6, Amp = 1.2, f = 0.54π, Mod = 0.25
10
0
0
50
DC = 19, Amp = 1.8, f = 0.84π, Mod = 0.18
20
0
100
0
-100
0
100
Time from Contact (ms)
Nature Neuroscience: doi:10.1038/nn.2283
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
DC = 8.6, Amp = 0.7, f = 0.36π, Mod = 0.16
8
0
-100
0
100
Time from Peak of
EMG (ms)
Phase in Whisk Cycle (radians)
Peak = 99, Baseline = 11
Peak = 132, Baseline = 13
Peak = 94, Baseline = 13
Peak = 109, Baseline = 28
Peak = 60, Baseline = 14
B2 Trial Averaged Touch
Response
40
0
100
1
-1
0
1
E2 Touch Response Versus Velocity of
Whisking Motion
DC = 103, Amp = 54, f = -0.58π, Mod = 1.0
0 100 200
DC = 48, Amp = 22, f = 0.11π, Mod = 0.92
0
100
Spike Rate (hz)
DC = 45, Amp = 24, f = -0.32π, Mod = 1.1
0
50 100
-100
0
100
Time from Contact (ms)
0.5
0
-0.5
1
0
−1
0.5
0
-0.5
0.5
0
-0.5
0.5
0
-0.5
-100
0
100
Time from Contact (ms)
Spike Rate (hz)
DC = 46, P2T = 43, PkVel = -0.68, Mod = 0.93
0
50 100
DC = 47, P2T = 26, PkVel = -0.22, Mod* = 0.55
0
100
DC = 104, P2T = 64, PkVel = 0.72, Mod = 0.62
0 100 200
DC = 137, P2T = 61, PkVel = 0.31, Mod* = 0.45
0 200
DC = 96, P2T = 91, PkVel = −0.56, Mod = 0.70
0
200
DC = 108, P2T = 68, PkVel = -0.6, Mod* = 0.63
0
200
D2 Touch Response Versus Phase in
Whisk Cycle (from video)
200
200
C2 Touch Response versus phase in
Whisk cycle (from EMG)
DC = 110, Amp = 57, f = 0.29π, Mod = 1.0
0
DC = 61, P2T = 25 PkVel = 0.7, Mod* = 0.41
0
100
200
DC = 62, Amp = 44, ph = 0.13, Mod = 1.4
0
100
DC = 107, Amp = 41, f = 0.13π, Mod = 0.77
0
DC = 64, Amp = 43, Φ = 0.06π, Mod = 1.3
0
100
-π
0
π
-π
DC = 96, Amp = 60, f = 0.32π, Mod = 1.3
0
0
200
100
100
DC = 138, Amp = 67, f = -0.17π, Mod = 0.97
0
200
0
DC = 49, Amp = 23, f = 0.17π, Mod = 0.94
0
DC = 104, Amp = 59, f = -0.35π, Mod = 1.1
0
100 200
DC = 135, Amp = 60, f = -0.28π, Mod = 0.89
0
200
−1
DC = 94, Amp = 58, f = 0.20π, Mod = 1.2
0
Velocity of Vibrissa Motion (deg / ms)
π
-π
0
π
-π
0
π
-π
0
π
-π
0
π
DC = 47, Amp = 17, f = -0.57π, Mod = 0.72
0
50
Spike Rate (hz)
200
DC = 45, P2T = 11, Ang = 132, Mod* = 0.24
0
50 100
DC = 48, P2T = 15, Ang = 138, Mod* = 0.31
0
100
DC = 101, P2T = 28, Ang = 98, Mod* = 0.28
0 100 200
DC = 136, P2T = 58, Ang = 145, Mod* = 0.43
0
200
DC = 94, Amp = 32, Ang = 98, Mod* = 0.34
0
DC = 108, P2T = 26, Ang = 106, Mod* = 0.24
0
200
DC = 62, P2T = 36, Ang = 155, Mod* = 0.58
0
100
F2 Touch Response Versus Vibrissal
Position
100
120
140
160
60
100
60
80
100
120
100
140
100
120
140
100
120
140
100
120
140
Spike Rate (hz)
50
0
50
0
20
0
20
0
0
50
0
10
0
10
500 µs
−20 0 20
Time from Spikes
Figure S8 (8-14) - Curtis and Kleinfeld
-100
0
100
Time from Contact (ms)
100 µV
A2 Trial Averaged Free
Whisking Response
DC = 15, Amp = 1, Φ = 0.17π, Mod = 0.13
20
0
0
100
DC = 34, Amp = 3.8, f = 0.33π, Mod = 0.23
40
0
0
100
DC = 14, Amp = 1.7, f = 0.30π, Mod = 0.24
20
0
0
100
DC = 15, Amp = 1.7, f = -0.31π, Mod = 0.22
20
0
20
0
Peak = 48, Baseline = 15
Peak = 49, Baseline = 11
-π
0
π
-100
0
100
Time from Contact (ms)
Nature Neuroscience: doi:10.1038/nn.2283
DC = 13, Amp = 3.0, f = -0.38π, Mod = 0.47
0
0
40
DC = 15, Amp = 2.2, f = 0.32π, Mod = 0.29
20
0
DC = 11, Amp = 2.3, f = -0.38π, Mod = 0.42
40
0
-100
0
100
Time from Contact (ms)
Angle Relative to Head Centerline (deg)
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
20
10
0
-100
0
100
Time from Peak of
EMG (ms)
Phase in Whisk Cycle (radians)
10
Spike Rate (Hz)
Spike Rate (Hz)
A3 Trial Averaged Free
Whisking Response
Peak = 85, Baseline = 24
B3 Trial Averaged Touch
Response
0
50
100
DC = 22, Amp = 2.1, f = 0.20π, Mod = 0.19
20
0
Peak = 65, Baseline = 11
-π
0
π
-π
0
DC = 11, Amp = 1.8, f = -0.06π, Mod = 0.32
20
40
π
Peak = 72, Baseline = 7
Peak = 60, Baseline = 10
0
40
Peak = 72, Baseline = 10
Peak = 47, Baseline = 17
Peak = 77, Baseline = 15
-π
0
π
-π
0
π
-π
0
π
-π
0
π
-π
0
π
C3 Touch Response versus phase in D3 Touch Response Versus Phase in
Whisk cycle (from EMG)
Whisk Cycle (from video)
E3 Touch Response Versus Velocity of
Whisking Motion
0
0.5
-0.5
0.5
0
-0.5
1
0
-1
1
0
-1
0.5
0
-0.5
-1
1
0
-1
0.5
0
-0.5
-100
0
100
Time from Contact (ms)
Spike Rate (Hz)
DC = 78, P2T = 52, PkVel = -0.49, Mod = 0.66
0
100
DC = 49, P2T = 35, PkVel = -0.27, Mod = 0.71
0
100
DC = 72, P2T = 69, PkVel = -1.01, Mod = 0.95
0
100
DC = 60, P2T = 37, PkVel = 0.96, Mod = 0.62
0
100
DC = 76, P2T = 50, PkVel = -0.79, Mod = 0.66
0
100
DC = 69, P2T = 21, PkVel = -0.06, Mod* = 0.30
0
100
DC = 83, P2T = 45, PkVel = -0.1, Mod* = 0.54
0 100 200
200
DC = 73, Amp = 48, f = 0.45π, Mod = 1.3
0
100
DC = 59, Amp = 38, f = -0.36π, Mod = 1.3
0
100
DC = 76, Amp = 44, f = 0.22π, Mod = 1.2
0
100
DC = 69, Amp = 33, f = -0.15π, Mod = 0.96
0
100
DC = 85, Amp = 38, f = 0.03π, Mod = 0.89
0 100 200
DC = 87, Amp = 43, f = 0.18π, Mod = 0.99
0 100
DC = 68, Amp = 27, f = -0.12π, Mod = 0.8
0
100
100
DC = 75, Amp = 43, f = 0.23π, Mod = 1.1
0
100
DC = 58, Amp = 45, f = -0.29π, Mod = 1.6
0
DC = 75, Amp = 49, f = 0.43π, Mod = 1.3
0
100
DC = 51, Amp = 49, f = 0.12π, Mod = 1.9
0
100
100
Spike Rate (Hz)
DC = 48, Amp = 43, f = 0.21π, Mod = 1.8
0
-100
0
100
Time from Contact (ms)
DC = 76, Amp = 43, f = 0.18π, Mod = 1.1
0
100
Spike Rate (Hz)
DC = 78, Amp = 42, f = 0.17π, Mod = 1.1
0
100
-100
0
100
Time from Contact (ms)
DC = 74, P2T = 20, Ang = 123, Mod* = 0.27
0
100
DC = 70, P2T = 22, Ang = 132, Mod* = 0.31
0
100
DC = 84, P2T = 27, Ang = 110, Mod* = 0.32
0 100 200
F3 Touch Response Versus Vibrissal
Position
80
100
120
110
130
100
DC = 60, P2T = 20, Ang = 144, Mod* = 0.33
0
100
DC = 75, P2T = 23, Ang = 146, Mod* = 0.31
0
100
DC = 50, P2T = 28, Ang = 158, Mod = 0.56
0
100
Spike Rate (Hz)
DC = 75, P2T = 23, Ang = 118, Mod* = 0.31
0
100
-100
0
100
Time from Contact (ms)
50
0
20
0
10
0
20
0
40
0
50
0
100
500 µs 100 µV
−20 0 20
Time from Spikes (mx)
0
Figure S8 (15-21) - Curtis and Kleinfeld
140
120
160
140
120
160
130
100
140
120
120
Angle Relative to Head Centerline (deg)
10
0
0
40
80
DC = 7, Amp = 1.9, f = 0.07π, Mod = 0.54
8
0
10
0
DC = 10, Amp = 1.7, f = -0.33π, Mod = 0.34
0
50
0
40
80
-100
0
100
Time from Contact (ms)
Nature Neuroscience: doi:10.1038/nn.2283
DC = 9, Amp = 1.6, f = 0.48π, Mod = 0.36
10
0
10
0
DC = 12, Amp = 2.7, f = 0.14π, Mod = 0.45
0
DC = 15, Amp = 2.0, f = 0.12π, Mod = 0.27
20
0
40
0
10
-100
0
100
Time from Peak of
EMG (ms)
Velocity of Vibrissa Motion (deg / ms)
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Phase in Whisk Cycle (radians)
Spike Rate (Hz)
B4 Trial Averaged Touch
Response
DC = 25, Amp = 15, f = 0.65π, Mod = 1.2
0
40
C4 Touch Response versus phase in D4 Touch Response Versus Phase in
Whisk cycle (from EMG)
Whisk Cycle (from video)
DC = 23, Amp = 12 f = 0.48π, Mod = 1.0
0
40
100
DC = 37, Amp = 28, f = 0.18π, Mod = 1.5
0
100
DC = 42, Amp = 25, f = -0.68π, Mod = 1.2
0
100
DC = 45, Amp = 26, f = 0.12π, Mod = 1.2
0
100
Spike Rate (Hz)
DC = 67, Amp = 60, f = 0.13π, Mod = 1.8
0
100
-100
0
100
Time from Contact (ms)
1
0
Spike Rate (Hz)
DC = 68, P2T = 37, PkVel = -0.2, Mod = 0.54
0
100
DC = 44, P2T = 25, PkVel = 0.06, Mod* = 0.57
0
100
DC = 42, P2T = 33, PkVel = 0.82, Mod = 0.79
0
100
DC = 35, P2T = 25, PkVel = -1.1, Mod = 0.71
0
100
DC = 55, P2T = 30, PkVel = -1.0, Mod = 0.55
0
100
DC = 49, P2T = 35, PkVel = 0.51, Mod = 0.71
0
100
DC = 25, P2T = 14, PkVel = -1.2, Mod* = 0.56
0
40
E4 Touch Response Versus Velocity of
Whisking Motion
-1
0
0.5
-0.5
1
0
-1
0.5
0
-0.5
1
0
-1
0.5
0
-0.5
-100
0
100
Time from Contact (ms)
100
120
140
100
120
140
130
150
100
120
140
130
140
150
100
120
140
120
130
140
F4 Touch Response Versus Vibrissal
Position
DC = 24, P2T = 12, Ang = 87, Mod* = 0.50
0
40
DC = 48, P2T = 22, Ang = 97, Mod = 0.46
0
100
DC = 57 P2T = 21, Ang = 124, Mod* = 0.37
0
100
DC = 34, P2T = 18, Ang = 130, Mod = 0.53
0
100
DC = 40, P2T = 9, Ang = 132, Mod* = 0.23
0
100
DC = 46, P2T = 22, Ang = 139, Mod* = 0.48
0
100
Spike Rate (Hz)
DC = 69, P2T = 53, Ang = 142, Mod = 0.77
0
100
-100
0
100
Time from Contact (ms)
0
20
20
0
0
20
20
0
0
10
25
0
10
500 µs 100 µV
0
−20 0 20
Time from Spikes (ms)
Figure S8 (22-28) - Curtis and Kleinfeld
Vibrissa Angle (deg)
A4 Trial Averaged Free
Whisking Response
Peak = 25, Baseline = 7
-π
DC = 35, Amp = 21, f = 0.22π, Mod = 1.2
0
DC = 40, Amp = 27, f = -0.72π, Mod = 1.4
0
100
Spike Rate (Hz)
100
DC = 46, Amp = 33, f = 0.20π, Mod = 1.4
0
100
100
DC = 70, Amp = 50, f = 0.1π, Mod = 1.4
0
0
1
0
DC = 46, Amp = 21, f = -0.69π, Mod = 0.91
0
100
0
-π
0
π
-π
0
π
-π
0
π
-π
0
π
-π
0
π
-π
0
π
-100
Time from Contact (ms)
DC = 56, Amp = 35, f = 0.35π, Mod = 1.3
0
100
20
Peak = 35, Baseline = 11
Peak = 38, Baseline = 7
Peak = 45, Baseline = 9
100
Peak = 66, Baseline = 10
0
Time from Contact (ms)
-100
DC = 54, Amp = 39, f = 0.38π, Mod = 1.7
0
100
-1
40
40
Peak = 50, Baseline = 10
π
Peak = 47, Baseline = 9
DC = 47, Amp = 23, f = -0.87π, Mod = 0.98
0
100
Whisking Velocity (deg / ms)
0
DC = 7, Amp = 1.1, f = 0.45π Mod =0.31
8
0
0
40
DC = 8, Amp = 1.1, f = -0.66π Mod =0.28
10
0
0
40
DC = 8, Amp = 1.0, f = 0.29π Mod =0.25
10
0
0
40
DC = 12, Amp = 2.1, f = 0.20π Mod =0.35
10
0
10
0
DC = 8, Amp = 0.6, f = -0.82π Mod =0.15
0
10
0
DC = 10, Amp = 1.3, f = 0.28π Mod =0.26
0
0
60
DC = 7, Amp = 1.2, f = 0.07π Mod =0.34
10
0
-100
0
100
Time from Peak of
EMG (ms)
Nature Neuroscience: doi:10.1038/nn.2283
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Phase in Whisk Cycle (radians)
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S9. Free whisking and touch responses for all units with rapid response to touch that are
not modulated by phase in the whisk cycle. (A). Average spike response histograms in
response to free whisking centered on EMG peaks with sinusoid fits (green lines) and resultant
parameters defined in caption for figure S8. An asterisk denotes insignificant fit. (B) Average
spike responses to touch with BARS fits (green lines). Peak and baseline rates calculated from
smooth curve are listed above each plot. (C) Average touch response histograms parsed by
phase in the whisk cycle. (D) Plots of the peak values of the touch response from fits to each of
the eight intervals of the smooth curve touch responses in panel C. The parameters listed above
each plot were estimated by fitting a sinusoid to each curve and have the same designation as
in panel A. Dotted lines indicate 95 % confidence intervals for mean spike rate estimates. (E)
Autocorrelations and waveforms for each neuron.
Nature Neuroscience: doi:10.1038/nn.2283
Spike Rate (Hz)
A Trial Averaged Free
Whisking Response
0
40
DC = 8, Amp = 0.5, Φ* = 0.29π, mod = 0.13
10
0
DC = 13, Amp = 1.8, Φ = 0.36π, mod = 0.27
B Trial Averaged
Touch Response
Peak = 32, Baseline = 8
Peak = 57, Baseline = 14
-π
0
π
0
-π
40
20
10
π
100
π
0
-π
π
0
-π
π
0
-π
π
0
-π
π
0
-π
0
Peak = 45, Baseline = 10
0
20
DC = 6, Amp = 0.2, Φ* = 0.08π, mod = 0.07
40
10
40
0
0
DC = 12, Amp = 1.5, Φ = −0.94π, mod = 0.25
20
10
100
0
0
Time from Contact (ms)
-100
Peak = 50, Baseline = 11
Peak = 46, Baseline = 20
0
0
DC = 28, Amp = 0.8, Φ* = −0.76π, mod = 0.06 Peak = 105, Baseline = 29
40
0
50
DC = 15, Amp = 0.7, Φ = 0.48π, mod = 0.10
20
0
40
0
0
DC = 12, Amp = 0.5, Φ* = 0.25π, mod = 0.08
10
0
-100
0
100
Time from Peak of
EMG (ms)
0
100
C Touch Response
Parsed by Phase
in Whisk Cycle
-100
Time from Contact (ms)
D
Touch
Response
versus Phase
in Whisk Cycle
DC = 35, Amp = 4, Φ* = 0.22π, mod = 0.2
50
0
DC = 58, Amp = 7, Φ* = −0.56π, Mod = 0.24
100
0
DC = 25, Amp = 3, Φ* = -0.10π, Mod = 0.24
50
0
DC = 48, Amp = 11, Φ* = −0.90π, Mod = 0.46
100
0
DC = 104, Amp = 18,Φ* = −0.87π, Mod = 0.35
200
0
DC = 45, Amp = 6, Φ* = 0.51, Mod = 0.27
50
Phase in Whisk Cycle (radians)
0
π
0
DC = 50, Amp = 2, Φ* = 0.09π, Mod = 0.12
50
0
-π
10
0
50
0
10
0
20
0
50
0
50
0
50
0
500 µs 100 µV
E
Unit
Autocorrelations
and Waveforms
−20
0
20
Time from Spikes (ms)
Figure S9 - Curtis and Kleinfeld
Nature Neuroscience: doi:10.1038/nn.2283
Spike Rate (Hz)
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Spikes / bin
Peak = 24, Baseline = 6
Phase in Whisk Cycle (radians)
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S10. Comparisons of maximum touch response and modulation depth at different
preferred phases. (A) The spike rate for the maximum touch response as a function of preferred
phase for units with whisking modulated rapid touch responses. (B) The modulation depth of the
touch response plotted as a function of each unit’s preferred phase. Mean values across all
preferred phases are indicated in each plot by dark green line along with 95 % confidence band
in light green. There is no significant tendency for either maximum touch response amplitude or
modulation depth to be greater at any specific preferred phase relative to other preferred
phases.
Nature Neuroscience: doi:10.1038/nn.2283
Protraction
Protracted
Retraction
Maximum Spike Rate
of Touch Response (Hz)
Maximum Touch Response by Phase in Whisk Cycle
200
μ+2σ
μ
100
μ−2σ
0
μ+2σ
Modulation Depth
of Touch Response
1.6
1.2
μ
0.8
μ−2σ
0.4
0
−π
−π/2
0
π/2
π
Preferred Phase in Whisk Cycle, φwhisk (radians)
Nature Neuroscience: doi:10.1038/nn.2283
Figure S10 - Curtis and Kleinfeld
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S11. The preferred phase in the whisk cycle for contact estimated from ∇EMG data
versus videographic data. The slope of the best fit line is 0.92 (p < 0.001).
Nature Neuroscience: doi:10.1038/nn.2283
0
π/2
0
-π
π
-π/2
Preferred Phase During Touch,
ftouch (radians), from ∇EMG
Count
10
-π
0
-π/2
π/2
Preferred Phase During Touch,
φtouch (radians), from Videography
π 0 10
Count
Figure S11 - Curtis and Kleinfeld
Nature Neuroscience: doi:10.1038/nn.2283
Supplemental Material for NN-A26718-T
“Phase-to-rate transformations encode touch in cortical neurons of a scanning
sensorimotor system” by John Curtis and David Kleinfeld
Figure S12. Touch responses as a function of phase in the whisk cycle for changes in whisking
parameters. Shown are examples, including that for unit 3 shown in figure 6 of the main text.
Nature Neuroscience: doi:10.1038/nn.2283
0
Unit 5
All
100
0
100
Spike Rate (Hz)
DC = 85, Amp = 48, f = 0.41π, Mod = 1.1
-100
Time after contact (ms)
f: 4 - 9
DC = 84, Amp = 41, f = 0.43π, Mod = 1.0
f: 9 - 13
DC = 83, Amp = 47, f = 0.36π, Mod = 1.1
amp: 10 - 16
DC = 85, Amp = 35, f = 0.33π, Mod = 0.8
0
100
0
100 200
Spike Rate (Hz)
DC = 83, Amp = 49, f = 0.41π, Mod = 1.2
amp: 16 - 24
-100
Time after contact (ms)
Unit 3*
All
100
0
100
Spike Rate (Hz)
DC = 73, Amp = 48, f = 0.45π, Mod = 1.3
0
Time after contact (ms)
-100
f: 5 - 9
DC = 73, Amp = 45, ph = 0.37, Mod = 1.3
f: 9 - 13
DC = 72, Amp = 46., ph = 0.40, Mod = 1.3
DC = 74, Amp = 45, ph = 0.46, Mod = 1.2
amp: 4 - 12
amp: 12 - 25
100
0
100
Spike Rate (Hz)
DC = 72, Amp = 46, ph = 0.38, Mod = 1.3
0
Time after contact (ms)
-100
Unit 12
All
100
0 100 200
Spike Rate (Hz)
DC = 103, Amp = 54, f = -0.58π, Mod = 1.0
0
Time after contact (ms)
-100
f: 4 - 8
DC = 103, Amp = 41, f = -0.54π, Mod = 0.80
f: 8 - 13
DC = 102, Amp = 58, f = -0.68π, Mod = 1.1
DC = 103, Amp = 53, f = -0.59π, Mod = 1.0
amp: 7 - 14
amp: 14 - 28
100
0 100 200 300
Spike Rate (Hz)
DC = 104, Amp = 50, f = -0.67π, Mod = 1.0
0
Time after contact (ms)
-100
Unite 20
All
100
100
0
Spike Rate (Hz)
DC = 51, Amp = 49, f = 0.12π, Mod = 1.9
0
Time after contact (ms)
-100
f: 6 - 8
DC = 53, Amp = 47, f = 0.17π, Mod = 1.8
f: 8 - 12
DC = 49, Amp = 46, f = 0.10π, Mod = 1.9
amp: 9 - 24
DC = 48, Amp = 42, f = 0.13π, Mod = 1.8
amp: 24 - 37
100
0 100
Spike Rate (Hz)
DC = 52, Amp = 45, f = 0.13π, Mod = 1.7
0
Time after contact (ms)
-100
Unit 21
0
All
100
0
100
Spike Rate (Hz)
DC = 76, Amp = 43, f = 0.18π, Mod = 1.1
-100
Time after contact (ms)
f: 4 - 8
DC = 73, Amp = 33.89, ph = 0.22, Mod = 1.14
f: 8 - 12
DC = 75, Amp = 51.54, ph = 0.13, Mod = 1.40
DC = 78, Amp = 43.26, ph = 0.18, Mod = 1.08
amp: 6 - 14
100
amp: 14 - 29
0
Unit 26
All
0
100
0
100
Spike Rate (Hz)
DC = 42, Amp = 25, f = -0.68π, Mod = 1.2
-100
Time after contact (ms)
f: 8 - 12
DC = 41, Amp = 27, f = -0.65π, Mod = 1.3
f: 5 - 8
DC = 43, Amp = 26, f = -0.51π, Mod = 1.2
amp: 3 - 15
DC = 42, Amp = 19, f = -0.64π, Mod = 0.9
0
100
0
100
Spike Rate (Hz)
DC = 44, Amp = 33, f = -0.61π, Mod = 1.5
amp: 15 - 28
-100
Time after contact (ms)
Figure S12 - Curtis and Kleinfeld
0
100
Spike Rate (Hz)
DC = 77, Amp = 38.53, ph = 0.21, Mod = 1.23
-100
Time after contact (ms)
Nature Neuroscience: doi:10.1038/nn.2283
All trials
Bottom 50 % fwhisk
Top 50 % fwhisk
Botom 50 % θo
Top 50 % θo