Supplementary data for \"Spiking irregularity and frequency modulate the behavioral report of single-neuron stimulation

Neuron, Volume 81

Supplemental Information

Spiking Irregularity and Frequency

Modulate the Behavioral Report

of Single-Neuron Stimulation

Guy Doron, Moritz von Heimendahl, Peter Schlattmann, Arthur R. Houweling, and Michael

Brecht

Figure S1

0

17

0

17 18%

0

19

0

19 11%

0

11

0

11 0%

0

13

0

13 8%

−1 0 10

20

0

59 73%

Time (s) Response

I

Low frequency

Medium frequency

Highfrequency

Catch trials

Microstimulation trials

2 mV

100 ms

A B CS

ingl

e-ce

ll st

imul

atio

n tri

als

Figure S1. Example frequency experiment, Related to Figure 4

In frequency experiments, neurons were stimulated using nanostimulation of different

intensities (25%, 50% and 100% of maximal current) to induce low, medium and high

frequency spiking, respectively; duration was adjusted to keep the product of duration

and current constant.

(A) Recording of a RS neuron during, from top to bottom, nanostimulation trials, a no-

current-injection catch trial and microstimulation. Triangles indicate stimulation onset

and offset artifacts. In the microstimulation trace, artifacts were partially clipped.

(B) Action potential (ticks) raster plots and first lick responses (red squares) during the

different stimulation conditions. Only 20/59 microstimulation trials are shown. The

neuron was inhibited shortly after microstimulation. No activity is shown during

microstimulation (gray area) because it could not be measured.

(C) Response rates for each stimulation condition.

Figure S2A B C

0

38

0

38 21%

0

39

0

39 28%

0

39

0

39 23%

0

39

0

39 18%

−1 0 10

40

0

119 98%

ResponseTime (s)

Catch trials

Microstimulation trials

Sin

gle-

cell

stim

ulat

ion

trial

s

400 mspulse

200 mspulse

100 mspulse

5 mV

100 ms

I

Figure S2. Example number experiment, Related to Figure 4

In number experiments, neurons were stimulated using short, intermediate and long

nanostimulation (100 ms, 200 ms and 400 ms), at fixed current intensity, resulting in

spike numbers proportional to the step current duration. conventions as in Figure S1.

Only 40/119 microstimulation trials are shown.

Figure S3

10 30 100 3000

0.5

1

Frequency [Hz]

Res

pons

e P

rob.

10 30 100 3000

0.5

1

Frequency [Hz]

Late

ncy

[s]

1 10 1000

0.5

1

Number of spikes

Res

pons

e P

rob.

1 10 1000

0.5

1

Number of spikes

Late

ncy

[s]

Effect of spike frequency

Effect of spike number

A B

C D

p = 0.7 p = 0.8

p = 0.4 p = 0.3

Figure S3. Effects of spike frequency and number on behavioral responses in FS neurons, Related to Figure 4

(A) Frequency experiment: adjusted response probability as a function of nanostimulation

spike frequency for FS neurons (n = 11). Each dot represents a cell and a condition. The

regression model (line plot) shows no significant effect (p = 0.7).

(B) Response latency after FS neuron stimulation as a function of frequency; no

significant correlation (p = 0.8).

(C, D) As in (A, B) for the number experiment (n = 21). There was no significant effect

of spike number for either response probability or latency (p = 0.4 and p = 0.3,

respectively).

Figure S4

All w/o N w/o f w/o CV 0

3165

3170

3175

3180

3185

BIC

Figure S4. Goodness-of-fit of the statistical model, Related to Figure 5

The plot shows the goodness-of-fit for different selections of regressors from the set of

CVISI, frequency (f) and number (N). Goodness-of-fit is given as the Bayesian

information criterion (BIC), where lower values indicate a better fit. The full model

(‘All’) is used as a reference value (red dashed line). Leaving out N yields the best model.

Figure S5

0 10 20 300

2

4

6

8

10

Frequency [Hz]

Pre

stim

ulus

pow

er (m

V2 )Nanostimulation trials

HitsMisses

0 10 20 300

2

4

6

8

10

Frequency [Hz]

Catch trials

FPsCRs

* *

*

A B

Figure S5. Increased power in the low beta frequency range precedes correct responses

(A) LFP power spectrum from the last two seconds before stimulation onset, averaged

across nanostimulation trials and cells, for hits (blue line) and misses (green line). Bars

indicate standard errors of mean. Total power in the 12-18 Hz range was significantly

different between hits and misses (two-sided paired t-test, p = 0.04), with a trend for the

full 4-30 Hz range (p = 0.2).

(B) As in (A) for catch trial false positives (FPs) and correct rejections (CRs). Total

power in the 12-18 Hz range was significantly different between FPs and CRs (two-sided

paired t-test, p = 0.01), as well as for the 4-30 Hz range (p = 0.03).

Figure S6

A B

C D10 30 100 300

−0.5

0

0.5

1

10 30 100 300

−0.5

0

0.5

1

10 30 100 300

−0.5

0

0.5

1p

R −

pC

atch

10 30 100 300

−0.5

0

0.5

1

Frequency [Hz]

pR

− p

Cat

ch

Frequency [Hz]

Figure S6. Sensory effect sign varies with AP frequency in RS neurons but not in FS neurons, Related to Figure 5

(A) Difference between adjusted response probability pR and adjusted catch probability pc

as a function of nanostimulation spike frequency for RS neurons from the frequency

experiment.

(B) As in (A) for RS neurons from the unified dataset.

(C, D) As in (A, B) for FS neurons.

Table S1. Experiment types used in this study

Experiment Nanostim.

duration (ms)

# Cells # putative

excitatory cells

# putative

inhibitory cells

1. Spike Frequency 100/200/400 66 55 11

2. Spike Number I 100/200/400 137 119 18

3. Spike Number II 200/400/800 40 37 3

4. Spike Irregularity 400 74 62 12

5. Nature 2008 200 70 59 11

Table S2. Effects of Spike frequency and number on response

probability using logistic regression model

RS Neurons, leaving out Spike Number

Parameter Estimate p Value

Intercept 0.2078 0.3109

Spike Irregularity (CVISI) 0.2541 0.0102

Spike Frequency (Hz) -0.3419 0.0015

Catch rate probability 0.7401 <.0001

RS Neurons, leaving out Spike Irregularity

Intercept 0.3851 0.0450

Spike Frequency (Hz) -0.3131 0.0071

Spike Number -0.08866 0.3722

Catch rate probability 0.7453 <.0001

FS Neurons, leaving out Spike Number

Intercept -0.1694 0.7566

Spike Irregularity (CVISI) 0.6942 0.0082

Spike Frequency (Hz) -0.1200 0.6160

Catch rate probability 0.7767 <.0001

FS Neurons, leaving out Spike Irregularity

Intercept 0.4410 0.3473

Spike Frequency (Hz) -0.3441 0.1932

Spike Number 0.2255 0.3787

Catch rate probability 0.7908 <.0001