Exp Brain Res (1988) 69:245-259 Exp.erimental Brain Research 9 Springer-Verlag 1988 State dependent activity in monkey visual cortex II. Retinal and extraretinal factors in V4 P.E. Haenny 1, J. H. R. Maunsell 2, and P.H. Schiller Department of Cognitive and Brain Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA Summary. Responses were recorded from isolated neurons in the visual cortex of rhesus monkeys while they performed an orientation match to sample task. In each trial the animal was first cued with randomly selected orientation, and then presented with a sequence of gratings whose orientations were ran- domly selected. The animal was required to release a switch when it saw a grating that matched the cued orientation. For some recordings the animal was given a tactile cue by having it feel the orientation of a grooved plate that it could not see. In other experiments the cue orientation was presented visu- ally on the screen in front of the animal and then removed before the sequence of gratings was pre- sented. Using this task it was possible to determine if a neuron's response to a particular orientation was affected by whether or not it was the orientation for which the animal was looking. Over half the neurons examined in V4 (110/192) responded differently to the visual stimuli when the animal was cued to look for different orientations. For some neurons responses to all stimuli were strong when the animal was cued to look for a particular orientation, but weak when the same stimuli were presented in trials where the animal had been cued to look for another orientation. This type of sensitivity was found in neurons recorded while the animal was given a tactile cue, and also in other neurons tested when a visual cue was used, suggesting that the activity was not of direct sensory origin. In support of this, neurons in V4 were not strongly affected when the animal felt the grooved plate while not performing the orienta- tion matching task. The prevalence of behavioral effects that was found using the orientation matching Present addresses: ~ Neurologische Klinik, Universit~itsspital, Rfimistr. 100, CH-8091 Ztirich, Switzerland 2 Department of Physiology, Box 642, University of Rochester, Rochester, NY 14642, USA Offprint requests to: J. Maunsell (address see footnote) task suggests that extraretinal signals respresent a prominent component of the activity in V4 of the behaving monkey. Key words: Vision - Extrastriate cortex - Attention - Single-unit recording - Monkey Introduction The results of the preceding paper and those from other studies (Fischer and Boch 1980, 1981, 1985; Fischer et al. 1981; Moran and Desimone 1985) show that the visual responses of neurons in V4 of the rhesus monkey can be greatly modulated depending on the behavioral significance of a visual stimulus. These findings suggest that signals arising from sources other than the retina may represent an important aspect of neuronal activity in visual cortex. Understanding the extent and nature of these extraretinal inputs is likely to be an important step in understanding the function of visual cortex. The experiments described here were directed at further characterizing the effects of behavioral states on responses in V4. We were particularly interested in determining whether the responses of neurons in V4 can be modulated by information supplied through another sensory modality. We have found that signals originating in the somatosensory system can be found in a large fraction of the neurons in this area. Another outcome of these experiments is the indication that the representation of behavioral state is not restricted to the modulation of sensory responses. Instead, neurons in V4 appear able to encode abstract, task-specific information. Finally, we found that at least one type of visual discrimina- tion reveals few behavioral effects in V4, suggesting that this cortical area may contribute in different246 ways to various visual tasks. A preliminary report of some of these findings has been presented elsewhere (Haenny et al, 1984). Methods Recordings were made from two alert, behaving rhesus monkeys (Macaca mulatta) that weighed 3.5 and 5.0 kg at the start of training. Neither animal had been previously used for experi- ments. Each was taught to move between its cage and a primate chair, and remained in its cage except during training or recording sessions. Early in the training a scleral search coil (Judge et al. 1980) and a head bolt were implanted under sterile conditions and barbiturate anesthesia. The head bolt was used to stabilize the animal's head during training and recording, and the search coil was used to monitor eye position (Robinson 1963). During training and recording the animal's fluid intake Was controlled and he worked for a juice reward. Standard operant conditioning tech- niques were used for training. A computer (DEC PDP 11/34) delivered stimuli, monitored eye position and collected data. Action potentials were recorded with a resolution of 1 ms. All stimuli were presented in pseudorandom sequences. One animal was trained to perform three different tasks. These were a tactile-visual orientation match, a visual-visual orientation match, and a task that required no matching (described below). Recordings were made after the animal learned each task. The other animal was trained to perform a single task: a selection from simultaneously presented stimuli. The recording techniques used for these experiments were identical to those described in the preceding paper. Tactile-visual orientation match The tactile-visual match required the animal to respond to the appearance of a visual grating whose orientation matched that of a grooved plate that the animal could feel but not see. At the start of each trial the computer rotated the grooved plate to one of four orientations and turned on a small fixation spot in the center of a tangent screen 75 cm from the animal (Fig. 1). When the animal had fixated the spot and pressed the grooved plate with his hand to close a switch, a series of visual gratings was presented. High contrast (-85%),