Selectively attending to or associating positive reinforcement with a given stimulus yield similar effects on behavioral performance. Subjects show improved accuracy and shorter reaction times when asked to detect changes in stimuli at attended versus unattended locations or when given larger rewards for successful performance. We sought to determine whether these factors have separable effects on responses in monkey primary visual cortex (V1). Previous studies have reported behavioral modulation (suppression or enhancement) of V1 responses, with the proportion of cells exhibiting modulation varying with task difficulty. To dissociate the effect of directed attention from reward on visual responses we trained monkeys to discriminate changes in stimulus orientation while separately varying and cueing: (i) the location of the target stimulus, (ii) the difficulty of the orientation discrimination, and (iii) the amount of reward offered for correct completion of a trial. This design allows for identification of the factors that modulate visual encoding in V1 neurons. For example, a given cells responses might be modulated by reward size but unaffected by task difficulty. Monkeys initiated trials by grasping a touch bar and fixating a central point. After a delay, a visual cue appeared briefly behind fixation. This cue signaled both the difficulty of the upcoming orientation discrimination and the amount of liquid reward available for a correct response. A short time after cue presentation, an array of six test stimuli (oriented Gabor patches) appeared, three on each side of the fixation point. In one hemifield, a stimulus was positioned within the classical RF (cRF) of an isolated single unit at the units preferred orientation, two flanking stimuli were then positioned and oriented immediately outside the cRF along an axis parallel to the central stimulus orientation. The positions of the three test stimuli in the opposite hemifield were diagonal reflections of these locations. After a variable delay, there was a change in the orientation of one of the two central stimuli;monkeys were required to release the touch bar shortly after this change to earn a reward. To bias spatial attention, across blocks of correct trials, we varied the probability that this change would occur within the cRF or in the opposite hemifield. The monkeys behavioral performance showed significant main effects of both reward size and magnitude of orientation change (ANOVA, both p <0.01), suggesting that these factors both influenced the allocation of spatial attention. Preliminary results include neuronal responses modulated by difficulty, spatial attention, or/and reward size. A powerful approach to studying how a brain contributes to behavior, the ability to reversibly silence neurons in selected brain regions would be a powerful method. Molecular biological manipulations have shown this approach to be possible and useful in rodents. Getting these techniques to work in nonhuman primates has been difficult. Ivermectin sensitive chloride channels can be used to reversibly silence neurons in behaving animals. We made use of a recently developed ivermectin sensitive chloride channel derived from the human glycine channel and created a YFP fusion protein (iClGly-Venus). To determine whether this channel can induce reversible behavioral changes in monkeys, we injected lentivirus expressing the iClGly-Venus protein from the synapsin promoter into the primary visual cortex of a rhesus monkey. We covered a 5 x 8 mm region on the medial operculum of one hemisphere of primary visual cortex, a region representing the contralateral lower visual field about 5-7 degrees from the fovea. To create a permanent scotoma as a reference, the monkey was also given an aspiration lesion in the other hemisphere covering tissue corresponding to 3-5 degrees from the fovea in the lower visual field. The monkey was trained to fixate on a spot in the center of the visual field, and to make saccades when a faint target 15 percent contrast against the background appeared briefly somewhere on a square grid spanning -9 to +9 degrees vertically and horizontally (120 locations). The monkey detected >90% targets in the upper visual fields as indicated by correct saccades. A target appeared at each location at least 20 times in each session. The monkey virtually never detected targets within the lesion representation (<10% of correct saccades). On average across three ivermectin treatments on the 3rd day following treatment (subcutaneous injections, 5 - 7.5 mg/kg) there was a decrease in correct saccades in the treated quadrant. This was confirmed by a paired t-test comparing 16 locations in each of the two lower hemifields after point-by-point subtraction from baseline behavioral performance (p <0.01). Mean performance difference between lower quadrants was 9.5 %. There was a maximum decrease in performance of 20% in frequency of target detection at the location 5-6 degrees diagonally down from the fovea into the lower field corresponding to the treatment. We conclude that the ivermectin sensitive glycine channel is an effective means for inducing a reversible behavioral loss of function.
Showing the most recent 10 out of 18 publications
