Behnam Karami^1,2, Alvin Dinh¹, Reza Azadi¹, Arash Afraz¹; ¹Laboratory of Neuropsychology, National Institute of Mental Health, NIH, Bethesda, MD, USA, ²Donders Centre for Cognitive Neuroimaging, Donders Institute for Brain, Cognition and Behavior, Radboud University, Nijmegen, The Netherlands

The inferotemporal (IT) cortex is the final stage of visual processing in the primate ventral stream. While IT responses to visual objects have been extensively characterized, far less is known about how direct stimulation of IT neurons modulates perception. In previous work, we showed that behavioral detection of focal IT stimulation depends on the image shown to the monkey, with detection performance lowest when no image is present. We next investigated the neural mechanism underlying this image dependence. We considered two hypotheses: (1) recurrent dynamics within IT suppress perturbations when global activity is near baseline, effectively nullifying optogenetic drive in the no-image condition; or (2) neurons are activated by optogenetic pulses, but in the absence of concurrent visual input, the evoked activity fails to reach the detection threshold for downstream decoders. To enable optogenetic control, we injected pAAV-CaMKIIa-C1V1(t/t)-TS-EYFP into IT cortex, expressing an excitatory opsin in pyramidal neurons. A macaque monkey was trained to maintain fixation while viewing object images or a blank screen. Each trial lasted 1 s and was separated by a 200 ms interval. On 75% of trials, 200 ms optogenetic pulses of varying intensity were delivered 400-600 ms after onset via a 120 μm Doric optical fiber integrated with a 24-channel Plexon V-probe, enabling simultaneous stimulation and recording. Optogenetic stimulation produced similar increases in IT spiking for image and no-image trials, contradicting hypothesis (1). However, ROC analysis showed a variable true positive rate across images, with the lowest value in the no-image condition (74% vs. 92%), supporting hypothesis (2): detection depends on a downstream thresholding mechanism, and without visual drive, IT stimulation alone is less likely to cross it. At least under passive viewing, these results suggest that a global readout, rather than local recurrent control, gates the impact of selective IT perturbations.

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