1. The rostromesial agranular frontal cortex of macaque monkey (Macaca fuscata), traditionally defined as the supplementary motor area (SMA), was studied using various physiological techniques to delineate two different areas rostrocaudally. 2. Field and unitary responses to electrical stimulation of the primary motor cortex were distinct in the caudal part, but minimal or absent in the rostral part. Intracortical microstimulation readily evoked limb or orofacial movements in the caudal part, but only infrequently in the rostral part. Neuronal responses to visual stimuli prevailed in the rostral part, but somatosensory responses were rare. The opposite was true in the caudal part. 3. The rostral part, roughly corresponding to area 6a beta, was operationally defined as the presupplementary motor area (pre-SMA). The caudal part was redefined as the SMA proper. 4. Single-cell activity in the pre-SMA was quantitatively compared with that in the SMA proper in relation to a trained motor task. 5. Phasic responses to visual cue signals indicating the direction of forthcoming arm-reaching movement were more abundant in the pre-SMA. 6. Activity changes during the preparatory period, which lasted until the occurrence of the trigger signal for the reaching movement, were more frequent in the pre-SMA. 7. Phasic, movement-related activity was more frequent in the SMA, and its onset was often time locked to the movement onset. In the pre-SMA, the occurrences of response time locked to the movement-trigger signal were more frequent than in the SMA. 8. Among neurons in both areas, directional selectivity was found in all the cue, preparatory, and movement-related responses.

1. Single-unit activity in the cingulate cortex of the monkey was recorded during the performance of sensorially (visual, auditory, or tactile) triggered or self-paced forelimb key press movements. 2. Microelectrodes were inserted into the broad rostrocaudal expanse of the cingulate cortex, including the upper and lower banks of the cingulate sulcus and the hemispheric medial wall of the cingulate gyrus. 3. A total of 1,042 task-related neurons were examined, the majority of which were related to the execution of the key press movements. In greater than 60% of them, the movement-related activity preceded the activity in the distal flexor muscles. 4. The movement-related neurons were distributed, in two foci, in the posterior and anterior parts of the cingulate cortex, both including the upper and lower banks of the cingulate sulcus. The posterior focus was found to largely overlap the area projecting to the forelimb area of the primary motor cortex by the use of the horseradish peroxidase (HRP) method. 5. About 40% of the cingulate cortical neurons showed equimagnitude responses during the signal-triggered and self-paced movements. The neurons exhibiting a selective or differential response to the self-paced motor task were more frequently observed in the anterior than in the posterior cingulate cortex. 6. The long-lead type of changes in activity, ranging from 500 ms to 2 s, were observed mainly before the self-paced and, much less frequently, before the triggered movements. They were particularly abundant in the anterior cingulate cortex. 7. Only a few of the neurons showed activity time-locked to the onset of the sensory signals. 8. These observations indicate that the anterior and posterior parts of the cingulate cortex are distinct entities participating in the performance of limb movements, even if the movements are simple, such as those in this study.

The cholinergic pedunculopontine tegmental nucleus (PPTN) is one of the major ascending arousal systems in the brain stem and is linked to motor, limbic, and sensory systems. Based on previous studies, we hypothesized that PPTN would be related to the integrative control of movement, reinforcement, and performance of tasks in behaving animals. To investigate how PPTN contributes to the behavioral control, we analyzed the activity of PPTN neurons during visually guided saccade tasks in three monkeys in relation to saccade preparation, execution, reward, and performance of the task. During visually guided saccades, we observed saccade-related burst (26/70) and pause neurons (19/70), indicating that a subset of PPTN neurons are related to both saccade execution and fixation. Burst neurons exhibited greater selectivity for saccade direction than pause neurons. The preferred directions for both burst and pause neurons were not aligned with either horizontal or vertical axes, nor biased strongly in either the ipsilateral or the contralateral direction. The spatial representation of the saccade-related activity of PPTN neurons is different from other brain stem saccade systems and may therefore relay saccade-related activity from different areas. Increasing discharges were observed around reward onset in a subset of neurons (22/70). These neurons responded to the freely delivered rewards within ~140 ms. However, during the saccade task, the latencies of the responses around reward onset ranged between 100 ms before and 200 ms after the reward onset. These results suggest that the activity observed after appropriate saccade during the task may include response associated with reward. We found that the reaction time to the appearance of the fixation point (FP) was longer when the animal tended to fail in the ensuring task. This reaction time to FP appearance (RTFP) served as an index of motivation. The RTFP could be predicted by the neuronal activity of a subset of PPTN neurons (13/70) that varied their activity levels with task performance, discharging at a higher rate in successful versus error trials. A combination of responses related to saccade execution, reward delivery, and task performance was observed in PPTN neurons. We conclude from the multimodality of responses in PPTN neurons that PPTN may serve as an integrative interface between the various signals required for performing purposive behaviors.

Neurons in the superior colliculus (SC) are organized as maps of visual and motor space. The companion paper showed that muscimol injections into intermediate layers of the SC alter the trajectory of the movement and confirmed previously reported effects on latency, amplitude, and speed of saccades. In this paper we analyze the pattern of these deficits across the visual field by systematically comparing the magnitude of each deficit throughout a grid of targets covering a large fraction of the visual field. We also translate these deficits onto the SC map of the visual/movement fields to obtain a qualitative estimate of the extent of the deficit in the SC. We found a consistent pattern of substantially increased saccadic latency to targets in the contralateral visual hemifield, accompanied by slight and inconsistent increases and decreases for saccades to the ipsilateral hemifield. The initial and peak speed of saccades was reduced after the injection. The postinjection amplitude of the saccades were either hypometric or normometric, but rarely hypermetric. Although errors in the initial direction of the postinjection saccades were small, they consistently formed a simple pattern: an initial direction with minimal errors (a null direction) separating regions with clockwise and counterclockwise rotations of the initial direction. However, the null direction did not go through the center of the inactivated zone, as would be expected if the SC alone were determining saccade direction, e.g., with a population code. One hypothesis that can explain the misalignment of the null direction with the lesion site is that another system, acting in parallel with the SC, contributes to the determination of saccadic trajectory.