A Neural Implementation of Wald's Sequential Probability Ratio Test

The following supplementary movies are prepared for the journal publication

Kira, S., Yang, T., Shadlen, M.N. (2015) A neural implementation of Wald's sequential probability ratio test. Neuron doi: 10.1016/j.neuron.2015.01.007

Movie S1 [related to Figures 3 and 4] . Example trial (Tin choice) with neural recording.

The movie depicts a trial in which monkey E chose the left target (Tin) after viewing 7 shapes. The movie is played at half speed. Eye position is depicted by the yellow dot. After acquiring fixation at the central white dot, two red choice targets appear. The left target is in the RF of the neuron. After a delay, shapes appear sequentially until the monkey initiates its saccade. Action potentials can be heard as clicks. They are also displayed as tick marks above the graph, which shows the cumulative logLR associated with each new shape (horizontal white line segments), displaced by 200 ms relative to the spikes in order to compensate for the sensory delay (τs; see main text). They are thus aligned approximately to the spikes induced by each new shape. The neutral level of evidence (logLR = 0) is shown as a horizontal yellow line. The eye position and lower graph were not displayed to the monkey; nor were the audio clicks. Only the fixation spot, targets, and shapes were visible to the monkey. In this trial, the cumulative logLR meandered around zero for the first few shapes and then favored Tin. Reflecting this pattern, the firing rate increased with the rise in the cumulative logLR until the monkey chose Tin.

Movie S2 [related to Figures 3 and 4]. Example trial (Tout choice) with neural recording.

The recording is from the same LIP neuron as in Movie S1. In this trial, the cumulative logLR initially favored Tin but then reversed to favor Tout. Reflecting this pattern, the firing rate increased initially but then diminished until the monkey chose Tout.