Search

Decision making and action selection are influenced by the values of benefit, reward, cost, and punishment. Mapping of the positive and negative values of external events and actions occurs mainly via the discharge rates of neurons in the cerebral cortex, the amygdala, and the basal ganglia. However, it remains unclear how the reward values of external events and actions encoded in the basal ganglia are integrated into reward value-based control of limb-movement actions through the corticobasal ganglia loops. To address this issue, we investigated the activities of presumed projection neurons in the putamen of macaque monkeys performing a visually instructed GO-NOGO button-press task for large and small rewards. Regression analyses of neuronal discharge rates, actions, and reward values revealed three major categories of neurons. First, neurons activated during the preinstruction delay period were selective to either the GO(large reward)-NOGO(small reward) or NOGO(large reward)-GO(small reward) combinations, although the actions to be instructed were not predictable. Second, during the postinstruction epoch, GO and NOGO action-related activities were highly selective to reward size. The pre- and postinstruction activities of a large subset of neurons were also selective to cue position or GO-response direction. Third, neurons activated during both the pre- and postinstruction epochs were selective to both action and reward size. The results support the view that putamen neurons encode reward value and direction of actions, which may be a basis for mediating the processes leading from reward-value mapping to guiding ongoing actions toward their expected outcomes and directions.

Animals optimize behaviors by predicting future critical events based on histories of actions and their outcomes. When behavioral outcomes like reward and aversion are signaled by current external cues, actions are directed to acquire the reward and avoid the aversion. The basal ganglia are thought to be the brain locus for reward-based adaptive action planning and learning. To understand the role of striatum in coding outcomes of forthcoming behavioral responses, we addressed two specific questions. First, how are the histories of reward and aversion used for encoding forthcoming outcomes in the striatum during a series of instructed behavioral responses? Second, how are the behavioral responses and their instructed outcomes represented in the striatum? We recorded discharges of 163 presumed projection neurons in the striatum while monkeys performed a visually instructed lever-release task for reward, aversion, and sound outcomes, whose occurrences could be estimated by their histories. Before outcome instruction, discharge rates of a subset of neurons activated in this epoch showed positive or negative regression slopes with reward history (24/44), that is, to the number of trials since the last reward trial, which changed in parallel with reward probability of current trials. The history effect was also observed for the aversion outcome but in far fewer neurons (3/44). Once outcomes were instructed in the same task, neurons selectively encoded the outcomes before and after behavioral responses (reward, 46/70; aversion, 6/70; sound, 6/70). The history- and current instruction-based coding of forthcoming behavioral outcomes in the striatum might underlie outcome-oriented behavioral modulation.

To examine the role of basal ganglia-cortical circuits in movement initiation, we trained monkeys to make the same arm movements in two ways-in immediate reaction to a randomly timed external cue (cued movements) and also following a variable delay without an explicit initiation signal (self-timed movements). The two movement types were interleaved and balanced in overall timing to allow a direct comparison of activity before and during the movement. Posterior putaminal neurons generally had phasic, movement-related discharges that were comparable for cued and self-timed movements. On cued movements, neuronal activity increased sharply following cue onset. However, for self-timed movements, there was a slow build-up in activity that preceded the phasic discharge. This slow build-up was time-locked to movement and restricted to a narrow time window hundreds of milliseconds before movement. The difference in premovement activity between cued and self-timed trials was present before the earliest cue-onset times and was not related to any differences in the overall time-to-move between the two types of trials. These features suggest that activity evolving in the basal ganglia-cortical circuitry may drive the initiation of movements by increasing until an activity threshold is exceeded. The activity may increase abruptly in response to an external cue or gradually when the timing of movements is determined by the animals themselves rather than an external cue. In this view, small changes in activity that occur in advance of the much larger perimovement neuronal activity may be an important determinant of when movement occurs. In support of this hypothesis, we found that even for cued movements, faster reaction times were associated with slightly higher levels of activity hundreds of milliseconds before movement.

To analyze the organization of corticostriatal motor inputs, we examined the neuronal responses in the putamen (Put) to stimulation in the primary motor cortex (MI) and the supplementary motor area (SMA). Stimulating electrodes were chronically implanted in the distal and proximal parts of the forelimb representation of the MI and in the forelimb representation of the SMA in Japanese monkeys (Macaca fuscata). Stimulation in the MI and SMA evoked orthodromic spike discharges in both phasically active and tonically active Put neurons. The latency of excitation evoked by MI stimulation was shorter than that of excitation evoked by SMA stimulation. Neurons responding exclusively to MI stimulation (MI-recipient neurons) and those responding exclusively to SMA stimulation (SMA-recipient neurons) were distributed predominantly in the ventrolateral and dorsomedial portion of the caudal aspect of the Put, respectively. About 20% of the recorded neurons responded concurrently to stimulation in both the MI and SMA (MI + SMA-recipient neurons). These neurons were located in the intermediate zone between the MI- and SMA-recipient zones. More than half of the Put neurons responded to sensorimotor stimulation. Movements of the forelimb were readily elicited by microstimulation in the MI-recipient zone, less frequently in the MI + SMA-recipient zone, and rarely in the SMA-recipient zone. More detailed analysis of the somatotopic arrangement based on cortical inputs, sensorimotor responses, and microstimulation-evoked movements revealed that within the MI- and MI + SMA-recipient zones of the Put, neurons representing the distal part of the forelimb were located more ventrally than those representing the proximal part. No such somatotopy was clearly detected in the SMA-recipient zone. The present results indicate that corticostriatal inputs from the forelimb regions of the MI and SMA are largely segregated. On the other hand, convergent inputs from the MI and SMA were noted on single neurons located at the junction between the two input zones. In addition, the corticostriatal inputs from the forelimb region of the MI exhibited a distal to proximal somatotopic organization along the ventrodorsal axis of the Put.

We used psychometric techniques and neurophysiological recordings to study the role of the putamen in somesthetic perception. Four monkeys were trained to categorize the speed of moving tactile stimuli. Animals performed a task in which one of two target switches had to be pressed with the right hand to indicate whether the speed of probe movement across the glabrous skin of the left, restrained hand was low or high. During the task we recorded the activity of neurons in the putamen contralateral (right) and ipsilateral (left) to the stimulated hand. We found different types of neuronal responses, all present in the right and left putamen. Some neurons responded during the stimulus period, others responded during the hand-arm movement used to indicate categorization, and others responded during both of these periods. The responses of many neurons did not vary either with the speed of the stimuli or in relation to the categorization process. In contrast, neurons of a particular type responded differentially: their activity reflected whether stimulus speed was low or high. These differential responses occurred during the stimulus and hand-arm motion periods. A number of the nondifferential and differential neurons were studied when the same stimuli used in the categorization task were delivered passively. Few neurons with nondifferential discharges, and none of the differential neurons, responded in this condition. In a visually cued control task we studied the possibility that the differential responses were associated with the intention to press or with the trajectory of the hand to one of the target switches. In this condition, a light turned on instructed the animal which target switch to press for a reward. Very few neurons in both hemispheres maintained the differential responses observed during the categorization task. Those neurons that discharged selectively for low or high speeds were analyzed quantitatively to produce a measure comparable with the psychometric function. The thresholds of the resulting neurometric curves for the neuronal populations were very similar to the psychometric thresholds. The activity of a large fraction of these neurons could be used to accurately predict whether the stimulus speed was low or high. The results indicate that the putamen, both contralateral and ipsilateral to the stimulated hand, contains neurons that discharge in response to the somesthetic stimuli during the categorization task. Those neurons that respond irrespective of the stimulus speed appear to be involved in the general sensorimotor behavior of the animal during the execution of the task. The results suggest that the putamen may play a role in bimanual tasks. The recording of neurons in the right and left putamen whose activities correlate with the speed categories suggests that this region of the basal ganglia, in addition to its role in motor functions, is also involved in the animal's decision process.

1. We studied the physiology of the neuronal projection from the striatum to the external and internal segments of the globus pallidus (GPe and GPi, respectively) in macaque monkeys. The objective of the study was to answer the following specific questions. 1) Which classes of the electrophysiologically identified striate neurons project to GPe and GPi? 2) What kind of information is transferred from the striatum to GPe and GPi during learned movement? 3) What are the physiological actions of striate projection neurons on target neurons in GPe and GPi? 4) What is the spatial pattern of the striatopallidal projections? 2. Sequential arm and orofacial movements were used as behavioral tasks. Visual stimuli triggered a sequence of three flexions-extensions of the elbow joint across the target, and the click of a solenoid valve triggered repetitive licking movements. 3. Striatopallidal projection neurons were electrophysiologically identified by antidromic activation after focal stimulation of either GPe or GPi. Of two classes of striate neurons, tonically active neurons (TANs) with tonic spontaneous discharges (2-8 imp/s) and broad action potentials, and phasically active neurons (PANs) with a very low spontaneous discharge rate (< 0.5 imp/ s) and high-frequency discharges in relation to behavioral tasks, PANs were identified as the projection neurons to either GPe or GPi. In 325 TANs examined by stimulation of GPe or GPi, no neuron was activated antidromically, even in the case of TANs located in the close vicinity of PANs that were identified as striatopallidal projection neurons. 4. The physiologically identified projection neurons (52 cells) in the striatum exhibited either discharges related to movement (30 cells) or discharges related to preparation for movement (4 cells) during performance of learned motor tasks. The activities of the remaining 17 striatopallidal neurons either were not related to the behavioral tasks used or could not be characterized sufficiently in the tasks. However, all of the unidentified striatopallidal neurons were PANs, on the basis of the spontaneous discharge rate and the shape of the action potential. 5. PANs with movement-related activity and those with preparation for movement-related activity were antidromically activated from the globus pallidus (GP). Not only the PANs that show burst discharges specifically at the beginning of a sequence of movement but also PANs that show phasic discharges time-locked to each movement of a sequence were identified as putaminopallidal projection neurons. On the other hand, no neurons that showed responses to sensory stimulus were identified as putaminopallidal neurons. 6. The conduction velocities of the putaminopallidal axons were estimated at approximately 1 m/s on the basis of the latency of antidromic activation and conduction distance. The PANs with activity only at the beginning of a sequential movement were more frequently found to project to GPi than to GPe, whereas the PANs with burst activity at each movement were more frequently found to project to GPe than to GPi. Among the GPi-projecting PANs, neurons with initial activity only showed a tendency to have longer latencies of activation from GPi than neurons with activity time-locked to each movement. 7. The physiological action of the striatopallidal projection was examined by switching from recording to microstimulation after identification of striatopallidal projection neurons in the putamen while recording evoked field potentials or spike discharges of single GP neurons located where the electrical stimulation evoked antidromic activation of the striate neurons with the lowest threshold. A small majority of GP neurons that exhibited increase of discharges during motor tasks received facilitatory putaminopallidal influences, whereas the vast majority of GP neurons that exhibited decrease of discharges during motor tasks received suppressive putaminopallidal influences.

1. Responses to metabotropic glutamate receptor (mGluR) activation were investigated in acutely dissociated rat neostriatal (caudate putamen, CP) large cholinergic neurons, with the use of a nystatin-perforated patch-clamp technique. 2. Application of L-glutamate (Glu) in the presence of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) induced a slow inward current accompanied by a decrease in membrane conductance at a holding potential (VH) of -44 mV. 3. The ratio of the maximal amplitude of the slow inward metabotropic glutamate (mGlu) response to that of the ionotropic glutamate response for large cholinergic CP neurons was larger than that for the hoppocampal CA1, CA3, and dentate gyrus neurons the nucleus tractus solitarii neuron, cerebellar Purkinje neuron, and granule cell of the main olfactory bulb. The threshold and half-maximal effective concentration values of these mGlu responses were 10- to 30-fold lower than those of the respective ionotropic responses. 4. Specific agonists of the mGluR, quisqualate (QA), (+/-)-1-aminocyclopentane-trans-1,3-dicarboxylic acid (tACPD), (1S,3R)-1-aminocyclopentane-1,3-dicarboxylic acid [(1S,3R)-ACPD], and (2S,3S,4S)-alpha-carboxycyclopropyl-glycine [L-CCG-1), similarly induced slow inward currents at the same VH. The relative affinities of the mGlu agonists were QA > Glu > L-CCG-1 > (1S,3R)-ACPD. L-CCG-1 did not induce any current at concentrations < 10(-6) M. 5. DL-2-amino-3-phosphonopropionic acid or DL-2-amino-4-phosphonobutyric acid did not block the mGlu response, whereas (RS)-alpha-methyl-4-carboxyphenyl-glycine, a selective mGluR antagonist, partially reduced the mGlu response. 6. The reversal potential of the mGlu response was close to the K+ equilibrium potential, and it shifted by 56.4 mV for a 10-fold change in extracellular K+ concentration. In 90.6% of the neurons tested, the instantaneous current induced by hyperpolarizing voltage steps was markedly suppressed during the mGlu response. In 9.4% of the neurons, the currents elicited by step pulses showed a voltage-dependent slow relaxation that was not affected by mGluR activation. 7. Under the current-clamp mode, the slow afterhyperpolarization (AHP) following a spontaneous discharge was not affected by tACPD. The AHP current was not blocked under the voltage-clamp mode, either. 8. (1S,3R)-ACPD or Glu in the presence of CNQX elicited membrane depolarization accompanied by increased rate of action potentials under the current-clamp mode. Tetrodotoxin had no effect on the membrane depolarization. 9. These results indicate that the mGluR on large cholinergic CP neurons is mainly of the mGluR1 and/or mGluR5 type, and it plays a significant role in controlling the membrane potential by way of suppressing the leak K+ conductance.

1. The basal ganglia of primates receive somatosensory input carried largely by corticostriatal fibers. To determine whether map-transformations occur in this corticostriatal system, we investigated how electrophysiologically defined regions of the primary somatosensory cortex (SI) project to the striatum in the squirrel monkey (Saimiri sciureus). Receptive fields in the hand, mouth, and foot representations of cortical areas 3a, 3b, and 1 were mapped by multiunit recording; and small volumes of distinguishable anterograde tracers were injected into different body-part representations in single SI areas. 2. Analysis of labeled projections established that at least four types of systematic remapping occur in the primate corticostriatal system. 1) An area of cortex representing a single body part sends fibers that diverge to innervate multiple regions in the putamen, forming branching, patchy fields that are densest in the lateral putamen. The fields do not form elongated cylindrical forms; rather, they are nearly as extended mediolaterally as they are rostrocaudally. 2) Cortical regions representing hand, mouth, and foot send globally somatotopic, nonoverlapping projections to the putamen, but regions with closely related representations (such as those of the thumb and 5th finger in area 3b) send convergent, overlapping corticostriatal projections. The overlap is fairly precise in the caudal putamen, but in the rostral putamen the densest zones of the projections do not overlap. 3) Regions representing homologous body parts in different SI cortical areas send projections that converge in the putamen. This was true of paired projections from areas 3a and 3b, and from areas 3b and 1. Thus corticostriatal inputs representing distinct somatosensory submodalities can project to the same local regions within the striatum. Convergence is not always complete, however: in the rostral putamen of two cases comparing projections from areas 3a and 1, the densest zones of the projections did not overlap. 4) All projections from SI avoid striosomes and innervate discrete zones within the matrix. 3. These experiments demonstrate that the somatosensory representations of the body are reorganized as they are projected from SI to the somatosensory sector of the primate putamen. This remapping suggests that the striatal representation of the body may be functionally distinct from that of each area of SI. The patchy projections may provide a basis for redistribution of somatosensory information to discrete output systems in the basal ganglia. Transformations in the corticostriatal system could thus be designed for modulating different movement-related programs.

1. In this study, the movement-related activity of putamen neurons was investigated in behaving monkeys. The objective of the study was to examine whether the activity occurring in phase with body movements is directly related to the movement per se by encoding movement parameters or whether it is dependent on the circumstances in which the movement is performed. 2. Sensorially triggered arm movements were used as a behavioral task. A sequence of three visually triggered repetitive flexion-extensions of the elbow joint across the target were followed by the delivery of a juice reward. 3. There are two classes of putamen cells: type I, with tonic spontaneous discharges (2-7 Hz) and broad extracellularly recorded action potentials, and type II, with very low spontaneous discharge rate (less than 1 Hz). The movement-related phasic discharges occur exclusively in type II cells. 4. The movement-related activity of type II cells is classified into two contrasting types of cells: type IIa that exhibit burst discharges preceding the first movement of a sequence of repetitive arm or orofacial movements but that are almost inactive during succeeding movements, and type IIb that show movement-locked burst discharges with one-to-one correspondence. The somatotopic location of the cells was identified by microstimulation and/or sensory responses to passive somatosensory manipulation of the periphery. 5. The activities of type IIa cells occur with a short and fairly constant latency after the visual trigger stimulus and cease as soon as the sequence of the learned movements is initiated. In the condition in which the monkey attended to the visual trigger stimulus without initiating learned movements and waited for the delivery of juice reward at a fixed time after the stimulus, type IIa cells exhibited slight but consistent phasic discharges after the visual stimulus with short latency. This indicates that the type IIa cells have a visuomovement property. The type IIb cells, on the other hand, have a longer latency of activity after the visual trigger than type IIa cells and do not have the visuomovement property. 6. The type IIa cells change their activity pattern depending on whether the direction of initial movement is predictable before the trigger stimulus or not. 7. The activities of type IIa cells in the arm area of the putamen precede the electromyogram (EMG) of prime mover muscles by greater than 100 ms on average, whereas most type IIb cells are activated after the EMG during a learned arm-movement task.(ABSTRACT TRUNCATED AT 400 WORDS)

Sensorimotor response properties of neostriatal neurons were characterized in conjunction with assessments of the motor effects of intrastriatal microstimulation in unanesthetized rhesus monkeys. Neuronal activity and microexcitability were assessed at 250- to 500-micron intervals and, in some cases, at 25- to 100-micron intervals. The results are based on the functional characterization of 878 putamen and 224 caudate neurons and analysis of the effects of microstimulation at each of these recording sites. Recording/stimulation sites were located between stereotaxic planes A6 and A22 in 81 microelectrode tracks from three monkeys. A total of 443 (50.4%) putamen neurons showed discrete responses to the sensorimotor examination. Of neurons with sensorimotor responses, 232 (52.4%) showed increased rates of discharge in relation to both active and passive movements of specific body parts. An additional 193 (43.6%) cells increased their rates of discharge only during the monkey's active movements of specific body parts. The remaining 18 (4.0%) cells appeared to respond exclusively to passive somatosensory stimulation. The sensorimotor response areas of putamen neurons ranged in size from an entire limb to a single joint. Putamen neurons were somatotopically organized throughout the rostrocaudal extent of the nucleus. Neurons with sensorimotor response areas involving the leg were located in the dorsolateral putamen, those with orofacial representations were located ventromedially, and those with arm representations were located in an intermediate position. Microstimulation evoked discrete movements of individual body parts at 21.6% of the 878 putamen sites. Over 95% (181/190) of the effective sites were located within the central half of the rostrocaudal extent of the putamen, between stereotaxic planes A10 and A17. The pattern of somatotopic organization revealed by microstimulation was the same as that derived from sensorimotor response properties of putamen neurons. Moreover, a close correspondence was observed between the movements evoked from a given SMZ and the functional properties of local neurons. In contrast to the results obtained in the putamen, none of the 224 stimulation sites in the caudate nucleus was microexcitable, and only 17 (7.6%) of the caudate neurons had definable sensorimotor response properties. This is consistent with the view that the primate putamen, by virtue of its anatomic connections with the sensorimotor and premotor cortical fields, is more directly involved in motor functions, whereas the caudate nucleus, by virtue of its connections with cortical "association" areas, is involved in more complex behavioral functions.