0 100 200 300 400 500 600 700 800 900 Spike density (sp/s) 0 100 200 300 400 -300 -200 -100 0 100 Time from target onset (ms) Time from saccade onset (ms) 50 70 90 110 130 150 0 100 200 300 400 500 600 -120 -100 -80 -60 -40 -20 0 100 200 300 400 500 600 Time from target onset (ms) Time from saccade onset (ms) 0 100 200 300 400 500 600 700 800 900 Neuron 1 Neuron 2 Neuron 3 Neuron 1 Neuron 2 Neuron 3 Neuron 1 Neuron 2 Neuron 3 Unstable population vector Stable population vector 0.4 1 Population stability SC FEF 0 100 200 300 400 -300 -200 -100 0 100 Time from target onset (ms) Time from saccade onset (ms) 0.4 1 caudal SC rostral SC + Fixation Point Stimulus Eye Express Regular Gap Time (ms) 0 100 200 300 400 -300 -200 -100 0 100 Time from target onset (ms) Time from saccade onset (ms) 0 100 200 300 400 -300 -200 -100 0 100 Time from target onset (ms) Time from saccade onset (ms) Normalized spike density (sp/s) 0 0.2 0.4 0.6 0.8 1 0.4 1 Population stability 0 100 200 300 Time from target onset (ms) 0 100 200 300 Time from target onset (ms) 0.4 1 Population stability 0 100 200 300 400 500 600 Spike density (sp/s) rostral SC caudal SC 0 0.2 0.4 0.6 0.8 1 0.4 1 0 0.2 0.4 0.6 0.8 1 0 100 200 300 400 -300 -200 -100 0 100 Time from target onset (ms) Time from saccade onset (ms) Normalized spike density (sp/s) 0 0.2 0.4 0.6 0.8 1 SC FEF 1 2 3 + + Fixation Point Stimulus Eye . . . Time (ms) 0 100 200 300 400 -300 -200 -100 0 100 Time from target onset (ms) Time from saccade onset (ms) 0 1 2 3 0 Fano Factor Uday K. Jagadisan1,4 and Neeraj J. Gandhi1,2,3,4 Departments of Bioengineering1, Otolaryngology2, Neuroscience3, and the Center for Neural Basis of Cognition (CNBC)4, University of Pittsburgh, PA, 15213 Population temporal structure supplements the rate code during sensorimotor transformations Funded by: NIH Grant EY022854 0 100 200 300 400 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -300 -200 -100 0 100 0 100 200 300 400 -300 -200 -100 0 100 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Time from target onset (ms) Time from saccade onset (ms) Time from target onset (ms) Time from saccade onset (ms) Population stability Population stability R 1 (t) R 2 (t) . . . R n (t) R 1 (t+τ) R 2 (t+τ) . . . R n (t+τ) . . . . R(t) R(t+τ) . . . . R(t) R(t+τ) . R 1 (t) R 2 (t) . . . R n (t) R 1 (t+τ) R 2 (t+τ) . . . R n (t+τ) . . . . . . . . R 1 (t) R 2 (t) . . . R n (t) R 1 (t+τ) R 2 (t+τ) . . . R n (t+τ) . . . . . . . . - Neurons involved in sensorimotor processing produce bursts for both motor and non-motor events - The mechanism by which downstream structures discriminate between sensory/cognitive and movement-related inputs in unknown - We recorded from the frontal eye fields (FEF) and superior colliculus (SC) in monkeys performing saccade tasks to probe this phenomenon Delayed saccade task Mean population activity Sensory burst Motor burst Why does the “sensory” burst not produce a movement? The across trial reliability of spiking activity (measured by Fano Factor) does not explain the discrepancy. Up close view of the population reveals a difference in the pattern of activity The population code is temporally inconsistent during the visual burst and consistent during the motor burst Instantaneous neuron shuffle Temporal shuffle Temporal stability in rostral SC shows the inverse profile The visual burst is stable during “express”-like saccades 533.06 GG32 0.4 1 Population stability 0 100 200 300 400 -300 -200 -100 0 100 Time from saccade onset (ms) Time from target onset (ms) SC FEF SC+FEF Caudal SC plus rostral SC plus anti-RF 0 100 200 300 400 -300 -200 -100 0 100 Time from saccade onset (ms) Time from target onset (ms) 0.4 1 Population stability 0 60 90 30 -90 -60 -30 5 20 40 2 10 0 60 90 30 -90 -60 -30 5 20 40 2 10 Introduction n = 17, 1 monkey n = 57, 2 monkeys FEF SC Summary and Conclusions Acknowledgements: Joe McFerron, Gloria Foster, Alexandra Maxim Shuffle controls: the stabilty profile is specific to temporal structure in the real data Combining all neurons from SC into one population - contralateral, ipsilateral, and rostral SC - preserves the stability profile Temporal stability - dot product between time-separated population unit vectors Red dots - point image of target location on SC map Blue dots - location of neuron on SC based on stimulation-evoked vector Combining neurons from SC and FEF preserves the pattern of temporal stability Gap task Temporal stability hypothesis: Stable population activity, coupled with an increase in firing rate, is necessary for movement generation. The unstable visual burst could, in effect, increase the threshold for integration into a saccadic burst - The sensory burst in visuomovement neurons in SC and FEF is unstable compared to the premotor burst. Thus, saccade generation requires stable population ac- tivity along with an increase in firing rate. - The reduction in population stability during the visual burst is attributable to the specific activity pattern in the recorded population. - Neurons in rostral SC show the opposite pattern for large saccades - increased stability during the visual burst in caudal SC and decreased stability during the saccade. - Population temporal structure is stable when a saccade is triggered off the visual burst as in short latency sac- cades in the gap task. - For more on how temporal structure could be decod- ed, go to poster 533.14 (HH4). Rostral SC neurons are active during fixation and microsaccades and lower their activity for large saccades. They further reduce their drive during large movements by becoming unstable at the population level.