HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 94/1/640    most recent 01201.2004v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (13) Citing Articles via Google Scholar Google Scholar Articles by Hoshi, E. Articles by Tanji, J. Search for Related Content PubMed PubMed Citation Articles by Hoshi, E. Articles by Tanji, J.

Neurons in the Rostral Cingulate Motor Area Monitor Multiple Phases of Visuomotor Behavior With Modest Parametric Selectivity

¹Department of Physiology, Tohoku University School of Medicine, Sendai; and ²The Core Research for Evolutional Science and Technology Program, Japan Science and Technology Agency, Kawaguchi, Japan

Submitted 22 November 2004; accepted in final form 3 February 2005

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
We examined the cellular activity in the rostral cingulate motor area (CMAr) with respect to multiple behavioral factors that ranged from the retrieval and processing of associative visual signals to the planning and execution of instructed actions. We analyzed the neuronal activity in monkeys while they performed a behavioral task in which 2 visual instruction cues were given successively with an intervening delay. One cue instructed the location of the target to be reached; the other cue instructed which arm was to be used. After a second delay, the monkey received a motor-set cue to be prepared to initiate the motor task in accordance with instructions. Finally, after a GO signal, the monkey reached for the instructed target with the instructed arm. We found that the activity of neurons in the CMAr changed profoundly throughout the behavioral task, which suggested that the CMAr participated in each of the behavioral processing steps. However, the neuronal activity was only modestly selective for the spatial location of the visual signal. We also found that selectivity for the instructional information delivered with the signals (target location and arm use) was modest. Furthermore, during the motor-set and movement periods, few CMAr neurons exhibited selectivity for such motor parameters as the location of the target or the arm to be used. The abundance and robustness of the neuronal activity within the CMAr that reflected each step of the behavioral task and the modest selectivity of the same cells for sensorimotor parameters are strikingly different from the preponderance of selectivity that we have observed in other frontal areas. Based on these results, we propose that the CMAr participates in monitoring individual behavioral events to keep track of the progress of required behavioral tasks. On the other hand, CMAr activity during motor planning may reflect the emergence of a general intention for action.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Cingulate motor areas (CMAs) have been defined in the banks of the cingulate sulcus in the medial wall of the cortical hemispheres of primates (Dum and Strick 1991; Kaada et al. 1949; Luppino et al. 1991; Matelli et al. 1991; Picard and Strick 1996; Smith 1945; Vogt and Gabriel 1993; Ward 1948). In subhuman primates, each CMA has been subdivided further into a rostral (CMAr) and a caudal portion (CMAc), which includes the dorsal (CMAd) and ventral (CMAv) subregions (Dum and Strick 1993; He et al. 1995). In this report, we deal with the properties of neurons within the CMAr.

The CMAr has anatomical connections with structures that control arm movements, such as the primary motor cortex (Dum and Strick 1991; Leichnetz 1986; Morecraft and Van Hoesen 1992; Muakkassa and Strick 1979; Shima et al. 1991; Tokuno and Tanji 1993) and the cervical segments of the spinal cord (He et al. 1995; Hutchins et al. 1988), which is consistent with a role in controlling arm movements. In addition, the CMAr is connected to premotor areas (Barbas and Pandya 1987; Hatanaka et al. 2003; Luppino et al. 2003; Pandya et al. 1981), including the supplementary and presupplementary motor areas (SMA and pre-SMA, respectively; Luppino et al. 1993; Wang et al. 2001). Further, the CMAr receive inputs directly or indirectly from the orbital and dorsolateral prefrontal cortex (Bates and Goldman-Rakic 1993; Lu et al. 1994; Morecraft and Van Hoesen 1993; Petrides and Pandya 1999, 2002; Takada et al. 2004; Van Hoesen et al. 1993); this raises the possibility that the CMAr also participates in the planning of motor behavior and in the cognitive aspects of motor control. The CMAr receives direct or indirect inputs by the cingulate gyrus (areas 23 and 24) from a variety of sensory association and limbic areas, including the amygdala, parahippocampal gyrus, parietal association cortex, temporal pole, insular cortex, and claustrum (Amaral and Price 1984; Baleydier and Mauguiere 1980; Barbas et al. 1999; Morecraft and Van Hoesen 1998; Morecraft et al. 2004; Ongur and Price 2000; Selemon and Goldman-Rakic 1988; Van Hoesen et al. 1993; Vogt and Pandya 1987; Vogt et al. 1987); these connections implicate the CMAr in the processing of perceptual and limbic information. Collectively, the anatomical connectivity of the CMAr suggests that this area collects a wide range of limbic, cognitive, emotional, and memory-contingent information and converts this information into motor information that is used to plan and execute voluntary actions (Paus 2001).

Despite the wealth of anatomical studies, few studies have been carried out to characterize the neuronal activity within the CMAr during the performance of motor behavior. Shima et al. (1991) were the first to report that CMAr neurons were active before the initiation of either visually triggered or self-initiated arm movements. Akkal et al. (2002) reported recently that CMAr neurons were active primarily during the preparation and execution of a 2-sequence reaching movement. Isomura et al. (2003) identified CMAr neurons that were active selectively for either a GO or NO-GO response or that selectively anticipated either a color or spatial cue. Shima and Tanji (1998) studied CMAr activity while monkeys performed a reward-based selection task in which information for motor selection was limited to changes in the amount of the reward. They identified 4 types of CMAr neurons that were active in such a manner as to suggest that these cells linked the occurrence of reward reduction and the selection of an alternative movement. That report revealed an aspect of neuronal activity that was consistent with a role for the CMAr in selecting forthcoming actions based on information about a reward (cf. Procyk et al. 2000). This raises the question: What are the other aspects of the participation of the CMAr in the cognitive control of motor behavior when behavior requires the transformation of sensorimotor information for the selection and initiation of an action?

In a series of reports, we described the properties of the neuronal activity within prefrontal cortex and medial and lateral premotor areas while monkeys performed a behavioral task that required the detection of 2 visual cues, the retrieval of relevant information from the cues, and the integration of the 2 sets of information before an action was planned. We showed that neurons in the dorsal and ventral premotor cortex (Hoshi and Tanji 2000, 2002, 2004c), the dorsal and ventral parts of the dorsolateral prefrontal cortex (Hoshi and Tanji 2004a), and the SMA and pre-SMA (Hoshi and Tanji 2004b) were involved specifically in the detection of instruction cues, the retrieval of information about components of required actions, the integration of 2 sets of information, and the preparation and execution of actions. Considering the anatomical connectivity of these frontal cortical areas to the CMAr, it was of great interest to examine the neuronal activity within the CMAr of the same subjects performing the same task. We found that the properties of CMAr neurons differed from those of neurons in other areas of the cortex with respect to multiple aspects of behavior. A preliminary account of this study appeared in abstract form (Hoshi et al. 2003).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Animals and apparatus

We studied 2 male monkeys (Macaca fuscata, 8 kg) that were cared for in accordance with National Institutes of Health guidelines and the Guidelines for Animal Care and Use published by our institute. The 2 monkeys had been used in our previous studies (Hoshi and Tanji 2000, 2002, 2004a,b). During the experimental sessions, each monkey sat in a chair, and its head was restrained. We installed 2 touch pads (17 cm apart) in front of the chair, and a color monitor equipped with a touch-sensitive screen was placed in front of the monkey (30 cm from its eyes). Eye positions were monitored with an infrared eye-camera system (R-21C-AS, RMS, Hirosaki, Japan). Neuronal activity was recorded with glass-insulated Elgiloy-alloy microelectrodes (1–2 M at 333 Hz), which were inserted through the dura mater using a hydraulic microdrive (MO-81, Narishige, Tokyo, Japan). Single-unit potentials were amplified with a multichannel processor and were sorted using a multispike detector (MCP plus 8, MSD; Alpha Omega Engineering, Nazareth, Israel). Electromyographic (EMG) activity was recorded with silver wire electrodes. We monitored the following muscles bilaterally during task performance: the biceps and triceps brachii, deltoid (anterior, lateral, and posterior heads), trapezius, flexor and extensor carpi radialis, supraspinatus, infraspinatus, pectoralis major, rhomboid, and neck and paravertebral muscles. The EMG activity was amplified and digitized with an A/D converter and the digital values were stored in a laboratory computer. The TEMPO/Win system (Reflective Computing, St. Louis, MO) controlled the behavioral task and saved data for off-line analysis.

Behavioral task

The monkeys were trained to perform a target-reach movement by following 2 sets of instructions: one instruction indicated the target location and the other indicated the arm to be used in reaching for the target (Fig. 1A). After an intertrial interval of 3 s, the task commenced when the monkey placed a hand on each touch pad and gazed at a fixation point (FP; 1.2° in diameter) that appeared at the center of the touch-sensitive screen. If fixation was maintained for 1,200 ms, the monkey was given the first instruction (the first cue; 400-ms duration), which contained information about either the target location or which arm to use. A small, colored cue that was superimposed on the central FP indicated the type of instruction, i.e., whether the instruction was related to the target location or arm use. For Monkey 1, a green circle or red square indicated the instruction for arm use, whereas a blue circle or red cross indicated the instruction for target location. For Monkey 2, a green square and blue cross indicated the instruction for arm use and target location, respectively. A white square (8 x 8°) that appeared to the left or right of the FP and appeared at the same time as the colored cue indicated the laterality of arm use (for the arm use–related instruction) or target location (for the target-related instruction). If fixation was maintained for 1,200 ms during the subsequent delay period (first delay), the second instruction (the second cue; 400 ms) was given to complete the information for the subsequent action. Thereafter, if fixation was maintained for 1,200 ms during the second delay, squares appeared on each side of the FP (set cue; 1,000 ms), which instructed the monkey to prepare to reach for the target when the FP disappeared (the GO signal). If the monkey subsequently reached for the target with a reaction time of <1 s, it was rewarded with fruit juice 600 ms after touching the screen. Before the GO signal appeared, Monkey 1 was required to fixate on the FP for 800–1,200 ms. The order of appearance of the target and arm instructions was alternated in a block of 20 trials, and laterality was randomized within each block. A series of five 250-Hz tones after a reward signaled reversal of the order of instructions.

View larger version (34K): [in this window] [in a new window]   FIG. 1. Behavioral task and recording sites. A: temporal sequence of behavioral events. Top row: trial in which 2 instructions were given, i.e., which arm to use ("arm") and which target to be reached ("target"), in that order. Bottom row: trial in which the 2 instructions were given in the reverse order. B: top view of the surface of the right frontal cortex. Gray box indicates an area within which the rostral cingulate motor area (CMAr) was projected. Double arrowhead with a dotted line indicates the most rostral extent of the corpus callosum. Single arrowhead with a solid line represents the border between the supplementary and presupplementary motor areas (SMA and pre-SMA, respectively). C: coronal section through the CMAr along the dotted line labeled "C" in B. Gray area represent the extent of the CMAr. AS, arcuate sulcus; CeS, central sulcus; CiS, cingulate sulcus. Scale bars in B and C = 5 mm. D: distribution of task-related neurons and intracortical microstimulation (ICMS) effects in the CMAr of Monkey 2. Individual electrode tracks were reconstructed from the coronal sections. Sites where task-related neurons were found or where ICMS evoked forelimb movement were projected to a cortical map unfolded along the line passing through the middle of the gray matter (the white line in C). Letters, a, b, and c, on the unfolded map, correspond to those in the coronal section in C. Size of the filled circle indicates the number of task-related neurons at each site. Black horizontal bars indicate the recording point where no task-related neurons were found. Gray horizontal bar indicates the site where ICMS evoked forelimb movements. Double arrowhead indicates the most rostral extent of the corpus callosum. Arrowhead represents the border between the SMA and pre-SMA. Scale bar = 1 mm.

After completing the behavioral training, an acrylic recording chamber and head-fixation screws were implanted on the skull under aseptic conditions. Anesthesia was induced with ketamine hydrochloride [8 mg/kg, administered intramuscularly (im)] plus atropine sulfate, followed by pentobarbital sodium (30 mg/kg im). Antibiotics and analgesics were used to prevent postsurgical infection and pain. Initially, skull openings were made to enable recordings from the prefrontal, premotor, and primary motor cortex in the left hemisphere. After 1 yr of recording period, we performed the second operation (under the same aseptic conditions and anesthesia) to make another skull opening over the medial part of the right cranium, for recording from medial motor areas (i.e., the right SMA, pre-SMA, and CMAr). At this stage, we performed simultaneous recordings from these 2 separate chambers in the left and right hemispheres, using 2 microelectrodes inserted independently into the cortex with separate microdrives. After collecting the neuronal data, the monkeys were deeply anesthetized with pentobarbital and perfused through the heart with saline, followed by a fixative of 4% formaldehyde with 10%, then 20% of sucrose. After marking the locations of the recording chamber at known electrode coordinates, the brains were removed from the skull. Then it was sectioned at 50-µm intervals in the frontal plane on a freezing microtome. Sections were Nissl-stained for histological reconstruction of the neuronal recording sites using electrode traces and iron depositions made by passing a positive DC current through the tip of the microelectrode.

Data analysis

Definition of task-related neurons and 10 task periods. We sampled all neurons for which activity was recorded during at least 4 blocks of trials (i.e., 80 trials). To define neuronal activity as task-related, we initially divided the behavioral task into the following 6 phases: 1) Control: 200–700 ms after attaining fixation; 2) prefirst cue: the 500-ms period before the first cue appeared; 3) first cue and delay: from 100 ms after the first cue onset until the onset of the second cue; 4) second cue and delay: from 100 ms after onset of the second cue until onset of the set-cue phase; 5) set cue: from onset of the set cue until the GO signal appeared; and 6) movement: the 500-ms period around the time at which movement started. We classified a neuron as "task-related" if the distribution of the discharge rate (spikes/s) was significantly different in at least one of 8 trial types (ANOVA, P < 0.05, repeated over 8 types of trials with 8 sequences of the presentation of the first and second cue). For the statistical analysis and display, data for the 5 task events were aligned separately to the onsets of the first and second cues, the set cue, the GO signal, and the time at which the screen was touched. These data were analyzed separately before being merged at the midpoint of the first and second delays and at the set-cue phase (i.e., 600 ms after cue offset and 600 ms before the onset of the second cue or the set-cue phase, and 600 ms after the set-cue onset and 600 ms before the GO signal).

Subsequently, to analyze the statistics of the properties of neuronal activity, we divided the entire task into one control period (200–700 ms after attaining fixation) and 10 task periods, which were defined as follows. 1) Precue: 500-ms period before the onset of the first cue; 2) first cue: 100–500 ms after the onset of the first cue; 3) early first delay: 500–1,000 ms after the onset of the first cue; 4) late first delay: last 500 ms before the onset of the second cue; 5) second cue: 100–500 ms after the onset of the second cue; 6) early second delay: 500–1,000 ms after the onset of the second cue; 7) late second delay: last 500 ms before the onset of the set cue; 8) early set cue: 500-ms period after the onset of the set cue; 9) late set cue: 500 ms before the GO signal appeared; and 10) movement: 500-ms period before the screen was touched.

Statistical analysis using interspike intervals. To analyze neuronal activity at a high temporal resolution, we first calculated the instantaneous firing rate as the inverse of the interspike interval (inverse-ISI; 1-ms resolution). As the rate of neuronal discharge tended to follow a Poisson distribution, the inverse-ISI data were square root–transformed to stabilize the variance (Zar 1999).

To estimate whether neuronal activity reflected information contained in the first and/or second cue, we used a one-way ANOVA. We examined how well each of the following equations accounted for neuronal activity

(1)

(2)

(3)

In Eqs. 1–3, the firing rate index is for inverse-ISI data that were sampled every 10 ms; ₀ is the intercept; and a, b, and c are coefficients. The categorical factors for the first and second cue are the 4 instructions provided in the cues (right arm, right target, left arm, and left target). The categorical factors for the combination of the main factors are the 4 possible combinations of arm use and target location provided in the first and second cue. First, we calculated the probability (P value) that the coefficient of each equation equaled zero. We calculated P values for each 10-ms time point (i.e., for each bin) by creating an algorithm that was executed with commercially available software (MATLAB 6.5, The MathWorks, Natick, MA). We took P < 0.01 to be statistically significant. Thereafter, we calculated the sum of squares (SS) between groups and divided this value by the total SS to obtain the SS ratio. These SS values were obtained from ANOVA tables by creating an algorithm that was executed with commercial software (MATLAB 6.5). The SS ratio was analyzed for each 10-ms bin of data. The larger the SS ratio, the better the firing rate index equation (above) represented neuronal activity. Based on the analysis of probability and the SS ratio, we classified the neurons into 4 categories according to whether instantaneous activity was best and significantly represented by 1) the first cue, 2) the second cue, or 3) the combination of information in both cues, or whether 4) none of the regression coefficients was significantly different from zero. This classification was carried out for data in every 10-ms bin.

Linear model analysis of activity that followed the appearance of the second cue. For the neuronal activity recorded after the presentation of the second cue, we quantified the extent to which the activity represented selectivity for the second cue or the combination of the first and second cues. We used the following linear model to carry out an ANOVA

(4)

In Eq. 4, the firing rate index is the square root–transformed firing rate during the second delay period, ₀ is the intercept, and ₁ and ₂ are coefficients. The categorical factors for the second cue (CUE2) are the 4 instructions given by the second cue (right arm, right target, left arm, and left target). The categorical factors for the combination of both the first and second cues (COMBINATION) are the 4 possible combinations of the 2 instructions for arm use and target location. We classified neurons into the following 4 groups after calculating the probability that coefficients ₁ and ₂ were zero: 1) selective for the combination only (P₁₌₀ 0.01 and P₂₌₀ < 0.01); 2) selective for the second cue only (P₁₌₀ < 0.01 and P₂₌₀ 0.01); 3) selective for both the second cue and the combination (P₁₌₀ < 0.01 and P₂₌₀ < 0.01); and 4) nonselective (P₁₌₀ 0.01 and P₂₌₀ 0.01).

Quantification of selectivity for arm use and target location

To examine the extent to which individual neurons exhibited selectivity for the location of the target or for which arm to use during the set cue and movement periods, we used a multiple regression analysis based on the following model

(5)

Target location and arm use could be either left or right. The default values of the variables for right and left in Eq. 5 were 0 and 1, respectively. We calculated the values of slopes ₁ and ₂ by dividing the difference in activity in spikes/s by the dimensionless initial variable values (1 and 0) to assess selectivity for target location and arm use, respectively. Selectivity for the left target was greater if ₁ >0, whereas selectivity was greater for the left arm if ₂ >0. A merit of this analysis was that we could measure selectivity using the dimension of firing rate (spikes/s).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Sampling of neuronal activity in the rostral cingulate motor area

Before collecting data, we first mapped the somatotopic organization of the medial motor areas, including the SMA, pre-SMA, and CMA. For this purpose, we systematically examined the effects of intracortical microstimulation (ICMS; 11–44 pulses, 200 µs wide at 333 Hz, current = 5–50 µA) and studied neuronal responses to somatosensory stimuli applied by the experimenter (Luppino et al. 1991; Matsuzaka et al. 1992; Mitz and Wise 1987). We used the same electrodes for the ICMS as for unit recordings. In the banks of the cingulate sulcus extending 5 mm rostral to the border between the SMA and pre-SMA, arm movements were evoked with ICMS; caudal to this, hindlimb movements were evoked (He et al. 1995). At the vicinity of the movement-evoked sites, we observed neuronal responses to somatosensory stimuli applied by the experimenter: responses to manipulation of joints in the forelimb or cutaneous responses to brushing the hairy skin in the forelimb.

We recorded neuronal activity 7 mm rostral from the caudal end of the forelimb representation area of the rostral cingulate motor area (CMAr; Fig. 1, B and C). The most rostral recording sites corresponded to the location of anterior end of the genu of the corpus callosum. In this study, we included both banks of the cingulate sulcus (as shown in Fig. 1C) as the CMAr. The CMAr in our study thus included area 24c and area 6c (Dum and Strick 1991; Matelli et al. 1991; Shima et al. 1991; Vogt and Vogt 1919). Because we did not find obvious differences in response properties of neurons within the CMAr, we combined the data obtained in the dorsal and ventral banks of the cingulate sulcus and its fundus region.

We recorded a total of 280 task-related neurons within the CMAr of the right hemisphere of 2 monkeys (100 and 180 neurons in Monkeys 1 and 2, respectively; Fig. 1, B and C). The success rate for the behavioral task during recording was >96% for both monkeys. A majority of the task-related neurons was located within the dorsal bank of the cingulate sulcus and the fundus region of the same sulcus (Fig. 1D, n = 256), where the activity of nearly one third of the CMAr neurons that we encountered and monitored on-line was task related. We found task-related neurons less frequently within the medial portion of the ventral bank of the cingulate sulcus (Fig. 1D, n = 24), where the activity of one tenth of monitored neurons was task related.

Changes in neuronal activity throughout the behavioral task

To study the time course of general changes in neuronal activity, we analyzed how many CMAr neurons exhibited a change in activity during each of the 10 periods of the behavioral task (see METHODS for definitions of each period). For each task period, we applied a paired t-test for each of the 8 sequences of the first and second cues (paired t-test, = 0.05, corrected for 8 repeated analyses) compared with the control period. The results are summarized in Fig. 2. The thick line shows that, in general, the number of neurons exhibiting changes in activity increased consecutively from the beginning to the end of the task periods. The fraction of task-related neurons was 24% at the time of the appearance of the first cue, increased to 45% on the presentation of Cue 2, and increased subsequently during the set-cue period to 63% at the time of movement onset. This time course of neuronal activity within the CMAr was different from time courses reported previously for neurons within the SMA and pre-SMA (Hoshi and Tanji 2004b).

View larger version (24K): [in this window] [in a new window]   FIG. 2. Time course of changes in neuronal activity within the CMAr during the 10 periods of the behavioral task. Thick line indicates the fraction of neurons that exhibited a significant increase or decrease in activity compared with the control period. Thin solid and dotted lines represent the fractions of neurons that exhibited a significant increase or decrease in activity compared with the control period, respectively. Note that an individual neuron could exhibit an increase in activity during one trial and a decrease in activity during another trial, i.e., the sum of the fractions could exceed the total number of neurons. Gray areas indicate task periods during which visual cues were presented (from left to right: first, second, and set cue). Numbers at the top are the actual numbers of neurons that exhibited a significant change in activity.

Before the appearance of the first cue (i.e., during the precue period), 13% (n = 39) of neurons exhibited a significant change in activity compared with the control period (Fig. 2). It is likely that the anticipatory activity reflected behavioral rules and/or specific expectations for the appearance of instructions related to target location or arm use because both sets of instructions (for arm use and target location) were presented in a fixed order within each block of 20 trials. To examine this possibility, we applied a 2-sample t-test for activity within the precue period (factor: order of instructions). Of the 39 neurons that exhibited anticipatory activity, none exhibited a significant difference in activity in response to each type of instruction (t-test, = 0.01). Even when the level was set at 0.05, only 7 neurons approached the significance level (with a median difference of 3.9 spikes/s). These results indicated that behavioral rules or specific expectations of forthcoming cues were represented only minimally within the CMAr.

Activity after the first cue

After the appearance of the first cue, many CMAr neurons exhibited significant changes in activity compared with the control period (Fig. 2); specifically, 24% (n = 68) during the first cue period, 35% (n = 99) during the early delay period, and 40% (n = 114) during the late delay period. During this first phase, 143 neurons exhibited either an increase or a decrease in activity compared with the activity of the same cells during the control period. For most of these neurons (96%), modulation of activity also occurred during the other task phases, as in the example shown in Fig. 3. In total, the activity of 35, 122, 113, and 97 of these neurons was modulated during the precue, second cue and delay, set cue, and movement phase, respectively. The activity of the remaining 6 neurons (4%) was modulated exclusively during the first phase of the task.

View larger version (66K): [in this window] [in a new window]   FIG. 3. Example of visual responses of a CMAr neuron. Raster displays and spike density functions (SDFs) indicate neuronal activity in trials sorted according to 8 task schedules with different orders of presentation of the 2 instructional cues. In the raster displays, each row represents a trial, and each dot represents a discharge from the neuron. SDFs (smoothed by the Gaussian kernel method; = 20 ms; mean ± SE) are presented below the raster display. Ordinate represents the instantaneous firing rate (spikes/s). Raster displays and SDFs are aligned to the onset of the first cue, second cue, set cue, and the GO signal, which were merged at the midpoint of the first delay, second delay, and set-cue period. From left to right: 3 gray areas indicate the presentation of the first cue, second cue, and the set cue. Instruction provided by each cue is indicated above each panel (RA, right arm; LA, left arm; RT, right target; LT, left target). To the right of each panel, activity during the movement period (aligned to the screen touch, ) is presented. Tic marks on the horizontal axis are placed at 400-ms intervals. This neuron exhibited visual responses to the appearance of the first, second, and set cue.

To analyze systematically how information provided by the first cue was represented by the activity of CMAr neurons, we examined neuronal activity by means of a 2-way ANOVA ( = 0.01) with the type of instruction (arm use or target location) and the position of the cue (left or right) as categorical factors. The activity during each of the 3 task periods (first cue, early delay, and late delay periods) was analyzed separately. We analyzed neurons for which activity changed significantly compared with the control period (Fig. 2). The number and proportion of neurons that exhibited significant selectivity for either the position or instruction, or both, are displayed in Fig. 4 (top panels). Most neurons that exhibited a change in activity were not selective for either the position of the cue or the type of instruction (63, 69, and 65% of neurons during first cue, early delay, and late delay periods, respectively). Very few neurons were selective for either the position of the cue or the type of instruction. It is noteworthy that 2/3 of neurons that exhibited a change in activity were not selective for information provided by the first cue. Our data on CMAr neurons can be compared with the data of Hoshi and Tanji (2004b) on pre-SMA neuronal selectivity because both data sets were obtained from the same animals performing the same behavioral task. The pie charts at the bottom of Fig. 4 indicate that more pre-SMA neurons were selective for the cue position or for the instruction than were nonselective. The proportion of neurons that exhibited selectivity during these periods was significantly greater in the pre-SMA than in the CMAr (P < 0.0001, ² test). These results suggest that CMAr neurons are much less involved than pre-SMA neurons in specifying visuospatial information or in retrieving specific information provided by the first cue.

View larger version (27K): [in this window] [in a new window]   FIG. 4. Selectivity for cue position and type of instruction in the CMAr. Pie charts summarize the proportion of neurons classified into 4 categories using a 2-way ANOVA (factors: cue position and type of instruction). Actual number of neurons in each category is presented next to the corresponding portion of each pie chart, which is identified below. Cue position–selective neurons were significantly selective (P < 0.01) only for the position of the white square. Instruction-only neurons were significantly selective (P < 0.01) only for the instruction. Neurons classified as position- and instruction-selective were significantly selective (P < 0.01) for both main factors (position and instruction) or for the interaction between the factors. In the bottom panels, data for neurons within the pre-SMA (Hoshi and Tanji 2004b) are shown in the same format.

After the appearance of the second cue, 40% (n = 169) of neurons exhibited a change in activity compared with the control period (Fig. 2); specifically, 44% (n = 125), 45% (n = 128), and 42% (n = 118) of the neurons exhibited a change in activity during the presentation of the second cue, early delay period, and late delay period, respectively. Most of the neurons (95%) that exhibited a change in activity also exhibited a change in activity during other task phases (Fig. 3); specifically, 20% (n = 35), 72% (n = 122), 82% (n = 139), and 62% (n = 106) of the neurons exhibited a change in activity during the precue, first cue and delay, set cue, and movement phases, respectively. The activity of the remaining 9 neurons (5%) was modulated exclusively during the second phase of the task.

We then examined whether the neuronal responses to the second cue were selective with respect to the instructional information provided by that cue or to the combination of information provided by the first and second cues. We quantitatively analyzed the extent to which neuronal activity in the CMAr represented the second cue or the combination of the first and second cues by using the model in Eq. 4. The results of this analysis are summarized in Fig. 5 (top panels). Neurons that were significantly selective for the second cue (P₁₌₀ < 0.01) or both the first and second cues (P₂₌₀ < 0.01) constituted a small proportion of the population of neurons that exhibited changes in activity. It is remarkable that a vast majority of the second-phase activity was not selective to either the second cue or the combination of the first and second cues (75% of neurons during the early delay and 90% during the late delay period). Such a lack of selectivity is exemplified by the neuron in Fig. 3. A comparison of these data with data from a previous study on pre-SMA neurons (Hoshi and Tanji 2004b) revealed that neurons that were selective for the combination of the first and second cues were more abundant within the pre-SMA relative to the CMAr (pie charts at the bottom of Fig. 5). Similarly, more pre-SMA neurons than CMAr neurons exhibited selectivity for the second cue (P < 0.001, ² test).

View larger version (23K): [in this window] [in a new window]   FIG. 5. Second cue and combination-selective activity within the CMAr. Each pie chart summarizes the proportion of neurons classified into 4 categories using a linear model analysis (factors: second cue presentation and combination of first and second cues). Number of neurons in each category is presented next to each subcategory, which is identified below. Neurons selective for the second cue exhibited a significant change in activity that was associated (P < 0.01) with the second cue only, whereas combination neurons were significantly associated (P < 0.01) with only the combination of the first and second cues. Neurons classified as both second cue–and combination-selective were significantly associated (P < 0.01) with both of the main factors. In the bottom panels, data for neurons within the pre-SMA (Hoshi and Tanji 2004b) are shown in the same format.

View larger version (21K): [in this window] [in a new window]   FIG. 6. Time course of neuronal activity within the CMAr that selectively represented the first cue, second cue, or both cues. Bin-by-bin plots of selective activity expressed as the fraction of neurons that were significantly most selective for the first cue (black trace), the second cue (blue trace), and the combination of both the first and second cues (red trace) are shown. Green dots and line indicate the fraction of neurons that exhibited a significant increase or decrease in activity compared with the control period (redrawn from the data in Fig. 2). To the right of the panel, an additional display shows the same data aligned to the screen touch (). Tic marks on the horizontal axis are placed at 800-ms intervals.

Activity during the set-cue phase

During the set-cue phase, 174 CMAr neurons exhibited a significant change in activity (Fig. 2). The fraction of neurons in this category increased from the early half to the late half of the set-cue phase and constituted 62% of neurons that exhibited task-related activity. In most cases, the increase or decrease in activity during the set-cue phase was substantial, as can be seen in the example shown in Fig. 7. Despite the intensity of the change in activity during the set-cue phase, the activity of individual neurons was not influenced by either the first or second cue or the combination of both cues. This was a common trend for the majority of CMAr neurons. As can be seen in Fig. 6, the fraction of neurons exhibiting activity that was selective for either the first or second cue or the combination of both cues never exceeded 10% during the set-cue phase.

View larger version (75K): [in this window] [in a new window]   FIG. 7. Set-related activity of a CMAr neuron. This neuron was intensely active during the second delay and set-cue periods, regardless of which arm was used or the location of the target. Display format is as in Fig. 3.

View larger version (26K): [in this window] [in a new window]   FIG. 8. Arm use and target location selectivity of set- and movement-related activities in the CMAr. Pie charts summarize the proportion of neurons classified into 4 categories using a 3-way ANOVA (factors: arm to be used, target location, and order of the 2 instructional cues). Actual number of neurons in each category is shown next to each segment, and the segments are identified below. Arm use–selective neurons were significantly selective (P < 0.01) only for the arm use. Target location–selective neurons were significantly selective (P < 0.01) only for the target location. Neurons classified as both arm use–and target location–selective were significantly selective (P < 0.01) for both main factors (arm to be used and target location) or for the interaction between the factors. In the middle and bottom panels, data for neurons within the pre-SMA and SMA (Hoshi and Tanji 2004b) are shown in the same format.

View larger version (29K): [in this window] [in a new window]   FIG. 9. Quantitative analysis of arm use and target location representation for individual CMAr neurons during the set and movement periods. A: scatterplot of the regression slope of arm use (horizontal axis) vs. the regression slope of target location (vertical axis) for 174 CMAr neurons during the late set-cue period. B: cumulative fraction of the absolute regression slopes of arm use and target location for the same neurons. C: scatterplot of the regression slope of arm use (horizontal axis) vs. the regression slope of target location (vertical axis) for 179 CMAr neurons during the movement period. D: cumulative fraction of the absolute regression slopes of arm use and target location for the same neurons during the movement period. For comparison, data from the pre-SMA and SMA (Hoshi and Tanji 2004b) are plotted in B (target selectivity in pre-SMA) and D (arm selectivity in SMA).

The activity of 63% (n = 179) of task-related CMAr neurons changed significantly during the movement period (Fig. 2). An example of a neuron that exhibited a marked increase in activity during this period is presented in Fig. 10; this neuron is typical in that it exhibited modest selectivity with respect to target location or arm use. To systematically analyze how the selectivities for target location and arm use were represented in CMAr neurons, we performed a 3-way ANOVA (factors: target location, arm to be used, and order of instructions; = 0.01) using the data from movement-related neurons. Of neurons that exhibited a significant change in activity during the movement period 80% were not selective for target location or arm use (top right panel in Fig. 8; P > 0.01 for arm to be used, P > 0.01 for target location, and P > 0.01 for the interaction). This is in sharp contrast to the abundance of arm selectivity that exists within the SMA (bottom right panel). The selectivity of pre-SMA and SMA neurons was also significantly more abundant than was the case for CMAr neurons (right panels in Fig. 8; P < 0.0001, ² test).

View larger version (82K): [in this window] [in a new window]   FIG. 10. Movement-related activity of a CMAr neuron. Intense activity was observed during movements irrespective of arm use or target location. Display format is as in Fig. 3.

Activity time-locked to the reward delivery

A drop of liquid reward was delivered 600 ms after the animal touched the correct target with the instructed arm. We found a population of neurons exhibiting activity changes that were time-locked to the delivery of the reward. Twenty-one neurons, which were defined as reward-related, exhibited a peak of activity during a period ranging from 200 ms before to 1,000 ms after the reward delivery. An example of activity of a reward-related neuron is shown in Fig. 11A. The reward-related neurons were not included in the analysis of selectivity for parameters of the forelimb movement described above.

View larger version (29K): [in this window] [in a new window]   FIG. 11. Reward-related activity and population activity. A: reward-related activity of a CMAr neuron. Intense activity was observed after the delivery of liquid reward. Display format is as in Fig. 3. Right panel: inverted triangle indicates the time when the monkey touched the screen, and the gray line represents when the liquid reward was delivered. B: population activity of neurons having an activity peak during the set-cue period (N = 65), displayed with time-varying spike density functions (smoothed by the Gaussian kernel method; = 20 ms). C: population activity of neurons having activity peak during the movement period (N = 96). D: population activity of neurons having activity peak around the reward-delivery period (N = 21).

Activity profile of individual neurons across task phases

As mentioned earlier, the activity of individual CMAr neurons was often modulated during more than one of the task phases. We therefore quantified the number of task phases during which individual neurons exhibited a significant change in activity (Fig. 12). We found that the activity of 25% of CMAr neurons increased or decreased during one of the 5 task phases (precue, first phase, second phase, set cue, and movement). For the remainder of the neurons, the activity of individual neurons changed during 2 (25%), 3 (18%), 4 (24%), and 5 (9%) phases.

View larger version (13K): [in this window] [in a new window]   FIG. 12. Number of activity-modulated task phases. Each bar represents the number of neurons that exhibited a change in activity during 1, 2, 3, 4, or 5 of the whole task phases (i.e., 1: prefirst cue, 2: first cue and delay, 3: second cue and delay, 4: set cue, and 5: movement periods). These data were based on the data presented in Fig. 2. Actual number of neurons is shown above each bar graph.

View larger version (48K): [in this window] [in a new window]   FIG. 13. Distribution of activity peaks and population activity. A: pie chart summarizing the fraction of neurons that exhibited the highest (top) or lowest (bottom) discharge rate within each of the 5 task phases. Number next to each section indicates the actual count in each class, which is identified below. B: activity profile of 280 neurons across the entire task periods examined. Neuronal activity was displayed consecutively by arranging the time of occurrence of the peak (top panel) or trough (bottom panel) as the task progresses from left to right. Identity of the 10 task periods is shown in the bottom. Density of dots in each bar indicates the relative magnitude of activity (relative to maximum and minimum values of each neuron across the 10 task periods). Each horizontal row represents activity of a single neuron.

View larger version (61K): [in this window] [in a new window]   FIG. 14. Precue activity of a CMAr neuron. This neuron was intensely active before the appearance of the first cue. Activity decreased as the cues were presented. Display format is as in Fig. 3.

We sampled neuronal activity of the CMAr in an area 7 mm in the rostrocaudal direction. It was of interest to examine whether task-related neurons had a particular spatial distribution. Therefore we examined the rostrocaudal location of CMAr neurons according to a classification based on the occurrence of the activity peak within each of the 5 task phases. We measured the distance in the rostral direction of recording sites for individual neurons from the most caudal recording site in which we had identified task-related neurons. The results are summarized in Fig. 15. There was a trend for caudally located neurons to be active during the set-cue or movement phase of the task, whereas rostrally located neurons tended to be active more during the early phases of the task. To assess this trend statistically, we classified the task-related neurons into 2 classes: 1) neurons exhibiting peaks of activity during task phases that preceded the presentation of information about the target to be reached and the arm to be used (i.e., before the second cue; n = 66) and 2) neurons exhibiting peaks of activity during task phases that followed the presentation of information about future movement (i.e., after the second cue; n = 214). Neurons in the former category were located significantly more rostrally than neurons in the latter category (Kolmogorov–Smirnov test, KS = 0.2540, P = 0.0023).

View larger version (21K): [in this window] [in a new window]   FIG. 15. Spatial distribution of task-related neurons in the CMAr. Histograms summarize the spatial distribution of the numbers of neurons in 5 categories classified according to activity peaks during individual task phases. Task phase during which activity peaked is indicated above each histogram. Horizontal axis represents the distance in the rostral direction (mm) measured from the most caudal point at which task-related neurons were located.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

Location of task-related neurons in the rostral cingulate motor area

Most task-related neurons were located within the dorsal bank and fundus region of the cingulate sulcus (Fig. 1D); few task-related neurons were located within the medial part of the ventral bank of the cingulate sulcus (Fig. 1D). Previous studies of CMAr activity during a sequential reaching or button-pressing motor task revealed task-related activity primarily within the dorsal bank of the cingulate sulcus and the adjacent region in the fundus (Procyk et al. 2000; Shima et al. 1991), rather than in the ventral bank of the cingulate sulcus. By contrast, when reward scheduling was required, task-related neurons were identified within the ventral bank of the cingulate sulcus and within the limbic striatum (Shidara and Richmond 2002; Shidara et al. 1998). These reports and the findings of the present study suggest that there is functional segregation within the CMAr: the dorsal bank and fundus region appear to be involved in motor behavior, whereas the ventral bank is involved in reward-contingent or motivational participation. This view is consistent with reports showing prominent limbic cortical projections to the cortex in the ventral bank of the sulcus (Morecraft and Van Hoesen 1998) and with an anatomical study that reported region-selective projections of the cingulate banks to the dorsal sensorimotor striatum and the limbic parts of the striatum (Kunishio and Haber 1994). Therefore task-related CMAr neurons in the present study appear to reside within sensorimotor pathways rather than in limbic pathways.

We also found regional differences in the manner in which CMAr neurons were involved in the different phases of the behavioral task. Caudally located neurons tended to be active during later task phases, particularly when a forthcoming motor task was instructed (i.e., after the appearance of the second cue in the behavioral task used in the present study). By contrast, neurons in which the activity peaked during the presentation of the first cue or during the first delay period tended to be located more rostrally. Although these trends for regional selectivity were only partial, this finding is consistent with the hypothesis that the caudal CMAr participates more in processes that are contingent on the initiation or execution of movement, whereas the rostral CMAr may be involved more in receiving sensory instructions. This concurs with anatomical reports that neurons that project to the spinal cord and primary motor cortex are located mainly within the caudal part of the CMAr (Dum and Strick 1991; He et al. 1995; Shima et al. 1991).

Modest selectivity for visual instruction cues and for planning and executing actions

Although many CMAr neurons exhibited a robust change in activity during each phase of the behavioral task, it was remarkable that the neuronal activity was only modestly selective for any of the relevant behavioral factors. First, the changes in the neuronal activity in anticipation of the presentation of visual instruction cues were weakly selective for forthcoming cues or behavioral rules. Second, only a small portion of the neuronal responses to the 2 instruction cues was selective with respect to what was displayed or instructed. Finally, during motor planning and execution, few CMAr neurons were selective for the location of the target to be reached or the arm to be used.

The modest representation of spatial features of visual signals that we identified accords with the results of 2 previous studies (Akkal et al. 2002; Isomura et al. 2003), which revealed little selectivity of CMAr neurons for the locations of cues, and with a study showing that lesions of the cingulate cortex did not affect spatial memory (Murray et al. 1989). Further, a recent study showed that neurons in the anterior cingulate cortex generally had longer latency for representing specific information inherited by the cues than the orbital and ventral prefrontal cortex (Xiang and Brown 2004). All of these results suggest that the anterior cingulate cortex (area 24) is not profoundly involved in the process of extracting specific information from visual signals. This view is consistent with the fact that sensory and limbic inputs to the dorsal area 24 chiefly come indirectly through areas 23 or 24ab (Morecraft and Van Hoesen 1998; Morecraft et al. 2004; Van Hoesen et al. 1993; Vogt and Pandya 1987; Vogt et al. 1992).

A striking feature of the present study was the paucity of selective activity that reflected information instructed by the 2 cues. This lack of selective behavior is in sharp contrast to the wealth of information represented within the dorsal premotor cortex (Hoshi and Tanji 2000), the dorsal part of dorsolateral prefrontal cortex (Hoshi and Tanji 2004a), and the pre-SMA (Hoshi and Tanji 2004b). As for the modest representation of motor parameters, 2 previous reports (Akkal et al. 2002; Procyk et al. 2000) also described relatively modest representation of the direction or location of a target to be reached. A novel finding of the present study was that most CMAr neurons exhibited similar activity regardless of whether the monkeys planned to use the right or left arm to reach the target. Such effecter-independent activity was also apparent during the execution of movement. Therefore the results of the present study and those of previous reports reveal that there is relatively modest parametric representation of actions within the CMAr. Such a modest representation of movement parameters also occurs within the claustrum, where 70% of movement-related neurons in the dorsal claustrum were reported to be nonselective for 3 different movements (pushing, pulling, or turning a manipulandum; Shima et al. 1996). The claustrum (Sherk 1986; Tanne-Gariepy et al. 2002) and the cingulate cortex (Baleydier and Mauguiere 1980; Van Hoesen et al. 1993) have reciprocal anatomical connections with widespread areas. This raises an intriguing possibility that an area that possesses highly converging inputs represents visuomotor parameters with less parametric selectivity, instead of representing varieties of specific information.

Monitoring of task events within the rostral cingulate motor area

What is the function of CMAr activity if selectivity for visual instructions and planned actions is modest? One possibility is that the CMAr signals the occurrence of the entire behavioral event to monitor the progress of a series of task phases. The successful completion of the behavioral task in the present study required the perception of visual signals, retrieval of necessary information, retention of this information, integration of 2 types of information, and the planning and execution of an action; therefore it is conceivable that an executive element was required to keep track of the multiple phases of the task. To this end, the existence of a neural structure that registers the occurrence of each behavioral epoch to monitor the progress of the phases of the task would be useful. Indeed, our findings are consistent with the notion that CMAr neurons can report the occurrence of a specific task event and might collectively monitor the progress of individual phases of the behavioral task.

The aforementioned view of the role of CMAr in regulating the task sequence by monitoring behavior is in line with previous reports that the CMAr represents the temporal relationship of multiple task events. For instance, CMAr neurons were reported to reflect the rank order of movements in serial reaching tasks (Akkal et al. 2002; Procyk et al. 2000