Alvin Dinh¹, Behnam Karami^1,2, Reza Azadi¹, Arash Afraz¹; ¹National Institute of Mental Health, ²Donders Institute for Brain, Cognition and Behavior

Visual processing is the result of complex interactions among neurons within and across visual areas. In inferotemporal (IT) cortex, neurons show a characteristic temporal response to visual stimuli: spiking activity rises, then declines, and settles into a plateau, typically described by a post-stimulus time histogram (PSTH). Two hypotheses have been proposed for the declining phase: (1) an intrinsic mechanism, where higher activation leads to fatigue and a drop in firing even in the presence of constant input; and (2) a network mechanism, where suppressive contributions from other neurons stabilize activity at a lower, more energy-efficient level. To directly test these hypotheses, we delivered 200 ms optogenetic pulses to a small region of macaque IT cortex transduced with pAAV-CaMKIIa-C1V1(t/t)-TS-EYFP to express an excitatory opsin in pyramidal neurons. We first measured each neuron’s typical PSTH in response to visual stimuli. We then trained an adaptive deep neural network model developed by Vinken et al., 2020, fitting its adaptation parameters in the Conv5 layer to match each neuron’s visual PSTH. Next, we tested whether these neuron-specific adaptation parameters would also reproduce the optogenetically evoked IT response curves. To do this, we stimulated Conv5 units with square-wave current injections matched to the magnitude of firing-rate modulation produced by optogenetic stimulation and computed the residual difference between the model’s prediction and the empirical optogenetically driven dynamics. We find that although at lower stimulation powers, neurons’ intrinsic adaptation parameters are sufficient to capture the artificial activation dynamics, they fail as the modulation level increases. These results indicate that intrinsic adaptation alone is not sufficient to explain the visual response dynamics; network interactions are also required, and their effects become increasingly apparent under stronger optogenetic drive.

Acknowledgements: This research was supported by the Intramural Research Program of the NIMH ZIAMH002958.