Shedding light on attentional control: a combined opto-fMRI-electrophysiology study in monkeys

We aimed to determine differences in top-down and bottom-up control of attention using hyperpolarizing optogenetics in monkeys. Macaques were trained to covertly detect a random dimming of a small target occurring with equal probability at one of four locations in each quadrant while distractors were presented in the other quadrants. Target location was cued either using a change in color of the target (bottom-up) or the fixation point (top-down, symbolic cue for each quadrant), respectively. After the cue, a delay period occurred during which the distractors dimmed, which had to be ignored by the animals. After a random delay period, the monkeys had to indicate the dimming of the target with a manual response in order to obtain a reward. Monkeys were trained to perform this covert spatial attention task while bottom-up and top-down cued trials were randomly presented. Consistent cue-dependent accuracy and reaction time effects were observed in unperturbed animals, as well as relatively small differential fMRI activations between bottom-up versus top-down cued trials in frontal, parietal and visual cortex (N =3, see Balan et al. SfN 2015). Guided by the fMRI maps, we injected an AAV2/5-CaMKII-Jaws-KGC-eGFP-ER2 viral vector in specific task-driven compartments of both LIP and FEF. After 6 weeks, we aimed to focally reduce neuronal activity in these task-driven foci. Optogenetic inactivation using red light (635nm) was focused on the cue period of randomly selected bottom-up or top-down cued trials, both during fMRI and electrophysiology. For light delivery, we used an optic fiber fixed to a microdrive either with (for electrophysiology and behavior) or without (for fMRI and behavior) electrode.Preliminary results from the LIP inactivation sessions showed mainly decreased but not completely abolished neural responses and decreased fMRI activity within the targeted section of LIP. During task performance, we also found a mixture of optogenetic-induced decreased and increased fMRI activity throughout nodes of the attention network. Furthermore, we observed decreased behavioral performance during short inactivation of the cue-period, which either affected the bottom-up or the top-down cued trials but rarely both together when the same location was inactivated. These results show that ultra-short reversible inactivation of LIP only during the cue period can affect top-down and bottom-up driven covert spatial attention behavior, as well as local activity and network dynamics.status: publishe