Wojciech Zajkowski¹, Katrina Quinn², Lenka Seillier³, Daniel A. Butts⁴, Hendrikje Nienborg¹; ¹NIH, ²University of Tubingen, ³Humboldt Foundation, ⁴University of Maryland

Microsaccades are small, involuntary eye movements that occur during visual fixation, and are thought to play a role in maintaining visual perception and preventing sensory adaptation. Microsaccades have long been linked to spatial attention, both as a consequence of attentional states, and as a potential mechanism for enhancing sensory processing. Here, we used statistical modeling to characterize the effects of attention and microsaccades on neural firing rates. We recorded extracellular single and multi-unit activity in early-to-mid visual areas (V1: 88 units, V2:1150, V3: 852, V4: 89) in 2 macaque monkeys, which performed a disparity discrimination task (113 sessions total), in which spatial attention was manipulated by presenting two stimuli simultaneously on both sides of the visual field, but only one of which was task-relevant. We modeled the neural spiking activity of each cell with a GLM, using stimulus and saccade onsets as predictors, and controlling for the effects of neural drift and within-trial adaptation. We show that, unlike prior reports, neural activity is affected by attention independently of microsaccadic modulation, even when considering the effects of microsaccade direction and correlations with attentional state. Across areas, a model where attention modulated stimulus processing had a higher log-likelihood for 98.6% of cells and explained 11% more of the maximal variance captured by the full model, than the model where attentional effects could only be implemented via direction-dependent saccadic modulation. Our results suggest that spatial attention modulates neuronal responses independently of saccades, consistent with an attention-dependent gain of the stimulus driven response. Moreover, these attentional effects are associated with an additive shift of the microsaccade kernels in early-to-mid visual areas.

Acknowledgements: This work was funded by the National Eye Institute Intramural grant 1ZIAEY000570-01 (H.N.) and NSF grant NCS-FO 2123568 (D.A.B.)