Figure 1.
Optogenetic stimulation is consistently detected after an initial learning period. A) Monkeys were first trained on a go/no go vibration detection task. Every trial began with an LED cue to place the hand on the touchpad. The monkeys task was to detect the vibration and respond by quickly removing his hand (<750 ms) to receive a reward. On approximately 50% of trials (Catch Trials), no vibration occurred and the monkey was rewarded for retaining its hand on the touchpad for at least 1.5 seconds. Detection performance was calculated using the proportion of responses on Stim trials (Hit rate) and the proportion of responses on catch trials (False Alarm Rate = Chance Performance). After learning the vibration detection task, the monkeys began the optical detection task to detect the optogenetic stimulation delivered directly into the somatosensory cortex (500 ms pulse of continuous stimulation). B) Example raster and peri-stimulus time histogram of multi-unit firing rate during the task. At t = −1 the monkey places its hand on the touchpad causing an increase in firing rate. On Stim Trials at t = 0, the stimulation comes on, increasing the firing rate and the monkey removes its hand if it feels the stimulation (example day with 25 correct trials in a row). On Catch Trials, no stimulation occurs and the monkey retains its hand on the touchpad (Removal during the response window indicates False Alarm).
Figure 2.
Learning to Detect Optogenetic Stimulation as Proxy for Mechanical Vibration.
A) Both monkeys required a learning period when switched to the optical detection task. Every block of trials included 100 trials with approximately 50% stimulation trials and 50% catch trials. Sensitivity index was used to calculate the detectability of the signal in order to compare sessions across blocks where chance performance varies. Peak performance in detecting mechanical vibration is shown as a reference to indicate that both monkeys learned to detect the optical stimulation approaching the performance level of the real physical vibration. Error bars represent the standard error of the mean (grey solid lines represent the sigmoid fit over experimental blocks). B) Summary of the overall detection performance for 500 ms light pulse stimulation for final 6 experimental blocks pooled for each monkey (Monkey S: n = 201 Optical, n = 378 Catch, Monkey I: n = 299 Optical, n = 301 Catch). Error bars are represent standard error of the mean for binomial parameter estimates. Distributions are significantly different by chi-squared (n = 310 and 337, p<<0.001 for both monkeys). C) Control experiment eliminating possible detection of inadvertent external stimuli. Neither neural nor behavioral modulation could be detected with the optical fiber disconnected. D) Control experiment for eliminating light induced heating effects. Stimulation in locations without opsin expression did not modulate neurons nor behavior and could not be detected. Error bars represent standard error of the mean.
Figure 3.
Varying Stimulus Duration and Power Intensity.
A) Varying stimulus duration between 10 and 200 ms shows that the optogenetic stimulation is detected with increasing accuracy as stimulus duration increases. Stimulus durations as short as 10 ms can be detected (Monkey S: n = 133 catch trials, and n = [19 30 19 23 19] trials for each duration [10 25 50 75 200] ms; Monkey I: n = 103 catch trials, and n = [25 15 19 21 24]). B) Increasing the power from 1.5 to 2.5 mW causes and increase in the detectability of the optical stimulation in both monkeys. Error bars indicate standard error of the mean. Data are fitted to a sigmoid (dashed lines).