Abstract

The ventral premotor cortex (PMv) has been implicated in the visual guidance of movement. To examine whether neuronal activity in the PMv is involved in controlling the direction of motion of a visual image of the hand or the actual movement of the hand, we trained a monkey to capture a target that was presented on a video display using the same side of its hand as was displayed on the video display. We found that PMv neurons predominantly exhibited premovement activity that reflected the image motion to be controlled, rather than the physical motion of the hand. We also found that the activity of half of such direction-selective PMv neurons depended on which side (left versus right) of the video image of the hand was used to capture the target. Furthermore, this selectivity for a portion of the hand was not affected by changing the starting position of the hand movement. These findings suggest that PMv neurons play a crucial role in determining which part of the body moves in which direction, at least under conditions in which a visual image of a limb is used to guide limb movements.

Introduction

Several lines of evidence suggest that the ventral premotor cortex (PMv) of primates is involved in the visual guidance of limb movements. Anatomical studies have revealed inputs from regions of the posterior parietal cortex that provide visuospatial information (Matelli et al., 1986; Dum and Strick, 1991; Kurata, 1991; Murata et al., 1997, 2000; Lacquaniti and Caminiti, 1998; Luppino et al., 1999). Lesion studies (Moll and Kuypers, 1977; Kurata and Hoshi, 1999; Fogassi et al., 2001) and studies in which cellular activity during the performance of visually guided forelimb movements was examined (Wise, 1985; Rizzolatti et al., 1988; Kurata, 1993; Hoshi and Tanji, 2002) have suggested that the PMv is involved in the visual guidance of limb movements when reaching for or manipulating objects. However, it is not known precisely which behavioral factor is represented in the activity of individual cells in the PMv during the visual guidance of motor behavior.

A recent report demonstrated that most PMv neurons that are tuned to the direction of movement encode movement information in an extrinsic frame of reference (Kakei et al., 2001). This finding revealed that the PMv is involved in encoding the direction of movement in space, rather than the direction of joint movement or the activity in the forelimb muscles that is required to achieve the limb movement. Although Kakei et al.'s report opened up a new avenue for studying the nature of neural representation in the PMv, it is still necessary to examine whether the reported PMv activity represents the direction of the moving hand itself, the direction of movement of the object that is acted upon, or the end point (target) of the movement. Therefore, in the present study we investigated whether the activity of PMv neurons encoded the direction of hand movement or whether the PMv neurons were concerned with controlling an image of the moving hand, under experimental conditions in which these two factors were dissociated. We show that the majority of PMv neuronal activity reflects the motion of the hand image being controlled rather than the movement of the physical hand or the target.

Materials and Methods

We trained a monkey (Macaca fuscata) to move its right hand while viewing an image of the same hand that was projected onto a video display unit after being captured with a video camera. The animal was cared for in accordance with National Institutes of Health guidelines and the Guiding Principles in the Care and the Guidelines for Animal Care and Use published by our institute. The motor task and devices that were employed in this study have been described in detail by us previously (Ochiai et al., 2002). The motor task was the capture of a target spot that was displayed on the video display unit with a specific part of the projected image of the hand; this was specified by the presentation of an instructional spot signal (Fig. 1A). The target and instructional spots were overlaid onto the video image of the hand. The task began when the animal placed its hand in the hold position that was displayed on the video display. If the hold was maintained for 1.0 s, a red spot was projected onto either the right or left side of the video image of the hand; this instructed the animal as to which side of the hand was to be used to capture the target (Hand-Cue; turned off after 1 s). Subsequently, after a delay period of 1.0 s, a green spot appeared at one of four target positions (Target-Cue). The animal was required to wait for 1.5 s after the appearance of the target spot until the initial hold position and the target cue had disappeared from the video display; this served as a ‘go’ signal to start the target-capturing movement within 500 ms. If the onset of movement occurred within 200 ms of the ‘go’ signal, the trial was aborted. The target cue was turned on when the animal reached the target with the instructed position of the hand image. The physical hand movement was either straight ahead or 30° to the right or left in the horizontal plane. The hand movement appeared either upward or to the right or left on the video display.

Figure 1.

(A) Temporal sequence of behavioral events. A monkey was required to capture a target (one of four spots) with the side of the hand that was cued by means of a spot that appeared on the video image of the animal's hand. (B) Alterations of starting position of the hand movement. The initial hold position of the hand was shifted to the right or left of the normal starting position by a distance that was equal to the distance between the targets that were to be reached. Such a shift in the starting position did not alter movement directions. (C) Cortical map of the recording sites. In the enlarged surface view (left), small dots indicate the points at which microelectrodes entered the brain, and large dots indicate penetration sites at which direction-selective neurons were recorded. The anteroposterior (AP) and mediolateral (ML) axes of the Horsley–Clarke stereotaxic coordinates are shown. Arc, arcuate sulcus; CS, central sulcus.

Figure 1.

(A) Temporal sequence of behavioral events. A monkey was required to capture a target (one of four spots) with the side of the hand that was cued by means of a spot that appeared on the video image of the animal's hand. (B) Alterations of starting position of the hand movement. The initial hold position of the hand was shifted to the right or left of the normal starting position by a distance that was equal to the distance between the targets that were to be reached. Such a shift in the starting position did not alter movement directions. (C) Cortical map of the recording sites. In the enlarged surface view (left), small dots indicate the points at which microelectrodes entered the brain, and large dots indicate penetration sites at which direction-selective neurons were recorded. The anteroposterior (AP) and mediolateral (ML) axes of the Horsley–Clarke stereotaxic coordinates are shown. Arc, arcuate sulcus; CS, central sulcus.

The monkey was also trained to perform the same motor task with the video image of the hand inverted horizontally. In this case, the motion of the video image of the hand was dissociated from the true motion of the hand. Under this inverted viewing condition, the video image of the hand was a mirror image of the visual image under the normal viewing condition; the hand image moved upward and to the left on the video display when the monkey moved its hand forward and to the right. After extensive training, the animal was able to adapt to this dissociation, and was able to capture the target without difficulty. After a transition period (10 trials under the inverted viewing condition), the animal's movement trajectories and muscle activities were indistinguishable from those that were observed under the normal viewing condition. Therefore, we analyzed the neuronal activity that was recorded after the initial 10 trials under the inverted viewing condition. In addition, the initial hold position of the hand was occasionally shifted to the right or left to examine the possible relation between neuronal activity and the location of the target (Fig. 1B).

We monitored eye positions at a sampling frequency of 250 Hz using an infrared corneal reflection monitor system (RMS Hirosaki). We monitored the trajectory of hand movement with an infrared position sensor (C5949 Hamamatsu Photonics). We also measured the activity of the arm and wrist muscles using electromyography. We used conventional electrophysiological techniques to obtain single-cell recordings from the monkey (Mushiake et al., 1991). An acrylic recording chamber was attached to the monkey's skull under aseptic conditions. During surgery, the monkey was anesthetized with ketamine hydrochloride (10 mg/kg i.m.) and pentobarbital sodium (30 mg/kg i.m.). We applied intracortical microstimulation with 10–20 pulses of 0.2 ms duration at 333 Hz. After the electrophysiological recording had been completed, histological examination of the recording sites was carried out in Nissl-stained coronal sections of brain. Electrode tracks were reconstructed with reference to microlesions that were made by passing direct current through Elgiloy alloy electrodes.

In the present study, we focused on premovement activity in the PMv and did not examine activity that was associated with visual responses to visual instructions or moving stimuli. We analyzed neuronal activity during the premovement period (200 ms following the GO signal) with reference to the activity during a control period (500 ms preceding the presentation of the initial hold position). If neuronal activity during the premovement period differed from that during the control period (Mann–Whitney U-test, P < 0.01), neuronal activity was defined as movement-related. Using this criterion, 154 PMv neurons were classified as movement-related neurons. Initially, we also analyzed neuronal activity during a waiting period (1 s preceding the GO signal), but because only 31 neurons were significantly active during this period (Mann–Whitney U-test, P < 0.01), we proceeded to examine only movement-related activity during the premovement period in the present study. We used a two-way analysis of variance (ANOVA) to determine whether movement-related neuronal activity reflected the direction of motion of the hand image or the part of the hand image that the animal was instructed to use to capture the target (right or left side of the hand). To evaluate the effects of changing the initial hold position of the hand, we used a three-way ANOVA, in which the direction of motion, side of the hand and initial hold position were used as factors. Thereafter, we compared premovement activity under the normal and inverted viewing condition to examine the effect of image inversion on neuronal activity.

Results

Behavioral Data

Prior to analyzing neuronal activity, we examined the trajectories of hand movements to determine whether these were influenced by the following behavioral factors: viewing condition (normal versus inverted), side of the hand used to capture the target (left or right), and the initial hold position. As shown in Figure 2, the trajectories of movement in three different directions were not affected by the aforementioned factors. We also analyzed the response time (between the onset of the ‘go’ signal and the onset of movement), peak velocity of movement, and activity of the forehand muscles. None of these variables was affected by the three behavioral factors (ANOVA, P > 0.05). We also analyzed the position and movement of the eyes, and failed to detect any influence on oculomotor behavior of image inversion or the side of the hand that was used to capture the target.

Figure 2.

Traces showing the trajectories of hand movements under three behavioral conditions: normal starting position under the normal viewing condition (left); normal starting position under the inverted viewing condition (middle); and right- or left-shifted starting position under the normal viewing condition (right).

Figure 2.

Traces showing the trajectories of hand movements under three behavioral conditions: normal starting position under the normal viewing condition (left); normal starting position under the inverted viewing condition (middle); and right- or left-shifted starting position under the normal viewing condition (right).

Directional Selectivity

We recorded 154 neurons exhibiting premovement activity within the PMv, posterior to the arcuate sulcus and lateral to the arcuate spur (Fig. 1C). Some recorded neurons were located within the upper portion of the posterior bank of the arcuate sulcus. We first investigated whether neuronal activity reflected movement direction. Among the 154 PMv neurons, 47 (31%; see Table 1) were found to be selective for the direction of the target-capturing movement (ANOVA, P < 0.001). In an example of a direction-selective PMv neuron (left column in Fig. 3), activity was most prominent when the monkey moved its hand leftward to capture a target. To investigate whether the directional selectivity of PMv neurons reflected the forthcoming direction of motion of the video image or the physical hand, we inverted the video image horizontally. For example, under this condition, rightward motion of the hand would appear to be leftward on the video display unit. In the example of the PMv neuron that is presented in Figure 3, when the video image was inverted horizontally, the neuron was highly active when the forthcoming motion of the image was in a left–forward direction (right column in Figure 3), despite the fact that the monkey was actually moving its hand in a right–forward direction. By comparing the activity in the two columns in Figure 3, it is obvious that the activity of this neuron reflected the forthcoming motion of the hand on the video display unit, rather than the motion of the physical hand. We tested the effect of image inversion on 35 direction-selective PMv neurons, and found that the activity of 23 (65%; see Table 2) direction-selective neurons was selective for image motion rather than the motion of the physical hand. Only two (6%) of the direction-selective PMv neurons were selective for the motion of the physical hand. For the remaining 10 (29%) neurons, we found that direction selectivity was lost under the inverted viewing condition.

Figure 3.

Discharges of a PMv neuron exhibiting selectivity for the direction of motion of the hand image to be controlled but not for the actual movement of the hand. The activity of this neuron was greatest when a hand movement was initiated for which the video image of the hand moved in a left–forward direction, regardless of whether the movement of the physical hand was left–forward (top left) or right–forward (top right). Histograms of the summed neuronal discharges in raster displays are aligned to the onset of movement. In this display, the animal was instructed to use the right side of the hand to capture a target. The bin width in the histograms is 100 ms.

Figure 3.

Discharges of a PMv neuron exhibiting selectivity for the direction of motion of the hand image to be controlled but not for the actual movement of the hand. The activity of this neuron was greatest when a hand movement was initiated for which the video image of the hand moved in a left–forward direction, regardless of whether the movement of the physical hand was left–forward (top left) or right–forward (top right). Histograms of the summed neuronal discharges in raster displays are aligned to the onset of movement. In this display, the animal was instructed to use the right side of the hand to capture a target. The bin width in the histograms is 100 ms.

Table 1

Classification of activity of PMv neurons during premovement period

Total
 
Direction selective
 
Hand-side selective
 
Both
 
Neither
 
154 (100%)
 
47 (31%)
 
67 (43%)
 
37 (24%)
 
77 (50%)
 
Total
 
Direction selective
 
Hand-side selective
 
Both
 
Neither
 
154 (100%)
 
47 (31%)
 
67 (43%)
 
37 (24%)
 
77 (50%)
 
Table 2

Distribution of motion selectivity revealed by image-inversion

 Numbers examined
 
Image motion
 
Hand-movement
 
 

 
 Selective
 
Selective
 
Non-selective
 
Direction-selective cells 35 (100%) 23 (65%) 2 (6%) 10 (29%) 
Hand-side-selective cells
 
39 (100%)
 
30 (77%)
 
0 (0%)
 
9 (23%)
 
 Numbers examined
 
Image motion
 
Hand-movement
 
 

 
 Selective
 
Selective
 
Non-selective
 
Direction-selective cells 35 (100%) 23 (65%) 2 (6%) 10 (29%) 
Hand-side-selective cells
 
39 (100%)
 
30 (77%)
 
0 (0%)
 
9 (23%)
 

Hand-side Selectivity

We subsequently investigated whether neuronal activity reflected which side of the video image of the hand was used to capture the target. We found that the activity of 67 (24%; see Table 1) of 154 movement-related PMv neurons differed according to which side of the hand the animal was instructed to use to capture the target. Of these neurons, 37 (37%; see Table 1) were found also to be direction-selective. Figures 3 and 4 illustrate a typical example of hand-side selectivity: the activity of a PMv neuron that was recorded during the capturing of a target using the right side of the video image of the hand (Fig. 3) is compared to the activity of the same neuron when the left side of the hand image was used to capture the target (Fig. 4). It is apparent that the PMv neuron displayed in Figures 3 and 4 was active selectively when the animal used the right side of its hand to capture the target. To further investigate whether such hand-side selectivity reflected selectivity for the anatomical features of the video image of the hand (such as selectivity for the radial or ulnar side of the hand) or the relative position (right or left side) of the hand image, we analyzed 39 hand-side-selective PMv neurons and compared the activity of these neurons under the normal and inverted viewing conditions. In this paradigm, the right side of the image of the hand corresponded to the ulnar side of the hand under the normal viewing condition, but corresponded to the radial side under the inverted viewing condition. In the example shown in Figures 3 and 4, neuronal activity was much greater when the animal was instructed to capture the target with the right side of the hand image, regardless of whether the image was inverted. Therefore, neuronal activity did not reflect selectivity for the anatomical position of the hand. We found that 30 (77%; see Table 2) neurons exhibited selectivity for the side of the hand image irrespective of whether the image was inverted; in the remaining nine neurons (23%; see Table 2), hand-side selectivity was lost when the image was inverted. Thus, selectivity for the side of the hand that the animal was instructed to use to capture the target was selective for the relative position of the hand in the video image, rather than for a particular anatomical feature or posture.

Figure 4.

Discharges of the same neuron shown in Figure 3 while the animal was capturing a target using the left side of the hand. The activity of this neuron was greatly diminished, irrespective of whether the animal was instructed to use the left side of the hand to capture the target under the normal (left column) or inverted viewing condition (right column).

Figure 4.

Discharges of the same neuron shown in Figure 3 while the animal was capturing a target using the left side of the hand. The activity of this neuron was greatly diminished, irrespective of whether the animal was instructed to use the left side of the hand to capture the target under the normal (left column) or inverted viewing condition (right column).

Effect of the Initial Hold Position

Our finding (see above) that the majority of PMv neurons that were selective for the forthcoming motion of the video image of the hand were also selective for the side of the hand in the video image that the animal was instructed to use to capture the target could be interpreted as indicating that the activity of these neurons reflected the spatial location of the target or the relative location of the movement trajectory with respect to the video screen. To rule out this possibility, we carried out an additional analysis under a behavioral condition in which the initial hold position of the hand was shifted to the right or left. An example of this analysis is presented in Figure 5. The neuron that was examined (the same neuron as in Figure 3) was highly active when the monkey captured a target with the right side of the hand, regardless of whether the initial hold position was shifted to the right or left (Fig. 5B,C). It is noteworthy that the activity of this neuron did not depend on the spatial position of the target. In Figure 6, the activity of the same neuron during the premovement period is replotted against either the direction of motion of the image (left panel) or target position (right panel) on the abscissa. For this neuron, activity clearly depended on the direction of motion of the image, irrespective of the target location. We performed the same analysis for 13 PMv neurons that exhibited selectivity for both the direction of movement of the video image and the side of the hand that the animal was instructed to use to capture the target. For nine such neurons (69%), neither type of selectivity was affected by changing the initial hold position of the hand image (P > 0.05). For the remaining four neurons (31%), the magnitude of selectivity either increased or decreased significantly when the initial hold position of the hand was shifted, but for all of these neurons, neither type of selectivity was altered. These results demonstrate that the selectivity of PMv neuronal activity did not reflect the location of the target and that neuronal activity did not depend on the relative position of the image of the hand on the video screen.

Figure 5.

Selectivity for the side of the hand that the animal was instructed to use to capture the target was not affected by a change in the starting position of the movement. An example of a PMv neuron (the same neuron shown in Fig. 3) that was selective for a left–forward direction of movement. (A) Activity of the neuron when the movement was started from the normal starting position. Activity was much greater when the right side of the hand image was used to capture the target. (B) Neuronal activity when the starting position of the hand was shifted to the right. Activity was relatively low when the left side of the hand was used to capture the same target (second left) that was captured with the right side of the hand from a normal starting position (top right panel). (C) Neuronal activity when the starting position was shifted to the left. Activity was intense even when the leftmost target was captured with the right side of the hand.

Figure 5.

Selectivity for the side of the hand that the animal was instructed to use to capture the target was not affected by a change in the starting position of the movement. An example of a PMv neuron (the same neuron shown in Fig. 3) that was selective for a left–forward direction of movement. (A) Activity of the neuron when the movement was started from the normal starting position. Activity was much greater when the right side of the hand image was used to capture the target. (B) Neuronal activity when the starting position of the hand was shifted to the right. Activity was relatively low when the left side of the hand was used to capture the same target (second left) that was captured with the right side of the hand from a normal starting position (top right panel). (C) Neuronal activity when the starting position was shifted to the left. Activity was intense even when the leftmost target was captured with the right side of the hand.

Figure 6.

Plotting discharges of a PMv neuron to show their dependence on the direction of image motion but on the position of targets. (A) The magnitude of the activity in the premovement period (ordinate, mean discharges/s) is plotted against the image-motion direction (abscissa). (B) Neuronal activity is plotted against the target position. Data for the three different start positions are indicated with three lines: red, black dot, and green for right-deviated start, normal start, and left-deviated start, respectively. L, leftward; U, upward, and R, rightward image motion.

Figure 6.

Plotting discharges of a PMv neuron to show their dependence on the direction of image motion but on the position of targets. (A) The magnitude of the activity in the premovement period (ordinate, mean discharges/s) is plotted against the image-motion direction (abscissa). (B) Neuronal activity is plotted against the target position. Data for the three different start positions are indicated with three lines: red, black dot, and green for right-deviated start, normal start, and left-deviated start, respectively. L, leftward; U, upward, and R, rightward image motion.

Examination of Population Activity

It seems of interest to know the extent to which the direction of image motion or the side of the moving hand was distinguishable with a population of PMv cells selective for either of the two behavioral factors. We first compared population activity observed in preferred and non-preferred directions of image motion, under normal and inverted-view conditions. For this purpose, we plotted peri-movement population activity based on averaging of peri-movement histograms constructed for individual neurons with 50 ms bin width. As shown in Figure 7a, the image direction was clearly distinguishable under normal, as well as under inverted view. The direction of actual hand-movement was distinguishable in only two neurons (Fig. 7b). Subsequently, we compared population activity, by plotting activity of neurons that were selective only for the hand side (Fig. 8a). Furthermore, we compared the direction selectivity of neurons that were selective only for the image-motion direction (Fig. 8b). In addition, we also compared population activity of neurons that were selective both for the image direction and hand side, under four behavioral conditions (Fig. 8c).

Figure 7.

Peri-event plots of population activity of PMv neurons showing direction selectivity under normal- and inverted-view conditions. (A) population activity of image-motion-selective cells. (B) Population activity of hand-movement-selective cells. Each data plot in the peri-event plot is calculated by averaging peri-event histograms of individual neurons aligned on the onset of movement. The data in histogram, with 50 ms bin, were normalized by taking a peak value as 1. In A, red arrows labeled a and b denote preferred and non-preferred image-motion directions. Under inverted-view conditions, the direction of actual hand movement to the direction labeled a′ with a black arrow will appear, in the image, to the direction indicated with a green solid arrow (labeled with a′). In B, red arrows labeled a and b mean preferred and non-preferred direction of actual hand movement.

Figure 7.

Peri-event plots of population activity of PMv neurons showing direction selectivity under normal- and inverted-view conditions. (A) population activity of image-motion-selective cells. (B) Population activity of hand-movement-selective cells. Each data plot in the peri-event plot is calculated by averaging peri-event histograms of individual neurons aligned on the onset of movement. The data in histogram, with 50 ms bin, were normalized by taking a peak value as 1. In A, red arrows labeled a and b denote preferred and non-preferred image-motion directions. Under inverted-view conditions, the direction of actual hand movement to the direction labeled a′ with a black arrow will appear, in the image, to the direction indicated with a green solid arrow (labeled with a′). In B, red arrows labeled a and b mean preferred and non-preferred direction of actual hand movement.

Figure 8.

Peri-event plots of population activity of PMv neurons belonging to three categories. (A) Neurons selective for hand side only. Comparison of activity for motions of preferred and non-preferred hand side. (B) Neurons selective for image direction only. Comparison of activity of preferred and non-preferred directions. (C) Neurons selective for both hand side and image direction. Display formats are basically the same as for Figure 7.

Figure 8.

Peri-event plots of population activity of PMv neurons belonging to three categories. (A) Neurons selective for hand side only. Comparison of activity for motions of preferred and non-preferred hand side. (B) Neurons selective for image direction only. Comparison of activity of preferred and non-preferred directions. (C) Neurons selective for both hand side and image direction. Display formats are basically the same as for Figure 7.

Comparison of Activity Properties among Neurons in PMv and PMd

As mentioned in a previous section, PMv neurons were active more often in the premovement period than during the delay period. In contrast, we reported previously (Ochiai et al., 2002) that PMd neurons were predominantly active during the delay period. Despite the difference of behavioral phases in which activity was observed, we examined similarities and differences in activity properties among PMv and PMd neurons. The population of neurons exhibiting selectivity for the image motion rather than actual hand-movement (tested by image inversion) was similar in PMv and PMd. On the other hand, in PMv the proportion of neurons that were selective solely for hand side (39%) was significantly more than that of neurons that were selective only for image motion (13%), whereas the proportions were indistinguishable in the PMd (Table 3).

Table 3

Distribution of direction-selective and hand-side-selective cells in PMv and PMd


 
PMv
 
PMd
 
Exclusively direction selective 10(13%)a 36(35%)b 
Exclusively hand-side selective 30(39%)a 31(30%)b 
Selective to both 37(48%) 36(35%) 
Subtotal
 
77
 
103
 

 
PMv
 
PMd
 
Exclusively direction selective 10(13%)a 36(35%)b 
Exclusively hand-side selective 30(39%)a 31(30%)b 
Selective to both 37(48%) 36(35%) 
Subtotal
 
77
 
103
 
a

P <0.001.

b

Not significant.

Discussion

In this study, we found that neuronal activity in the PMv during the premovement period of a target-capturing task in which an animal was guided by a visual image of its hand predominantly reflected the motion of the image of the hand to be controlled, rather than the motion of the physical hand. We also found that the activity during the premovement period of half of the direction-selective PMv neurons differed according to which part of the image of the hand the animal was instructed to use to capture the target. Such selectivity for a particular side of the hand was not affected by altering the initial position of the hand in the video screen. These findings suggest that the PMv is involved in controlling the motion of the visual image of the moving hand with reference to a starting point that was defined within the hand.

Neurons that were selective for forthcoming image motion were located within the rostral part of the PMv in the present study. Previous studies have reported that neuronal activity in the same area is related to motor acts that are performed with the forelimb (Weinrich et al., 1984; Wise, 1985). In other studies, neuronal activity in the PMv has been reported to reflect spatial target location (Godschalk et al., 1985; Gentilucci et al., 1988; Mushiake et al., 1997; Hoshi and Tanji, 2002), the three-dimensional shape of motor targets (Murata et al., 1997) and visual objects in the peripersonal space (Graziano et al., 1994; Fogassi et al., 1996). Our findings accord with the view proposed in these reports that PMv neurons predominantly reflect visual factors in spatial guidance of forelimb movements. The most salient property of PMv neurons we found was that the selectivity of the premovement activity corresponded to the visual coordinates created by the video system. This observation is in line with a report by Iriki et al. (2001) that the visual receptive fields of the polysensory neurons of the parietal cortex were transferred to the video screen, after training the monkey to grasp the food morsel guided by the video image. Kurata and Hoshi (2002) also described the representation of screen-based visual coordinates by PMv neurons, and reported further that the PMv is concerned with the transformation of visual coordinates created by prism adaptation of reaching. A previous report from our laboratory (Ochiai et al., 2002) also demonstrated that PMd neurons reflected the direction of image motion in the video system, in a similar fashion as PMv neurons, despite the fact that neuronal activity was observed predominantly during a delay period.

Our study has direct relevance to a recent study by Kakei et al. (2001) in which the activity of PMv neurons was examined in a monkey that performed wrist movements in four directions while maintaining three different postures. Kakei et al. reported for the first time that the activity of PMv neurons reflected the direction of wrist movement independent of forearm posture. This observation implied that direction selectivity was independent of the changes in the angles of joints or the combination of muscle activity that was required for each of the four directional movements. It is remarkable that the findings of Kakei et al. revealed that the activity of PMv neurons encoded the direction of movement in space. Nevertheless, as the authors pointed out in their paper, these findings may be interpreted in at least three different ways: PMv neurons may represent the direction of limb movement in space, the direction of movement of the object that is acted upon, and/or the goal or target of the movement. The findings of the present study appear to support the second of the three explanations, although we studied neuronal activity during the performance of a motor task that was dependent on the visual display of the moving hand, which differs from the conditions of the task that was used by Kakei et al. In our study, the hand image served as a visual object to be acted upon, equivalent to a cursor to be operated in the video screen. Alternatively, the animal might have recognized the hand image as a part of its own body. Either way, under this requirement, the PMv activity appeared not to be concerned with controlling movement in an intrinsic limb-based coordinate frame. Instead, it seems to be involved in controlling movement in a coordinate frame that adapted to the visual reference frame provided by the hand image projection system. Our finding is consistent with recent observations by Schwartz et al. (2004) demonstrating cortical activity as representing image motion to be controlled. The properties of our PMv activity, however, favor the interpretation of action rather than perception as a primary representation of the PMv.

In the dorsal part of the premotor cortex (PMd), Shen and Alexander (1997) reported neuronal activity reflecting the spatial target being reached for. At first sight, our present findings appear to contradict this report because PMv activity did not reflect the spatial location of the target on the screen (see Fig. 6). However, it is possible that neuronal activity in the PMv may reflect the spatial location of the target in a hand-centered frame of reference, rather than a body-centered frame of reference. Also, the target representation of Shen and Alexander might have been centered on a cursor that moved with the hand. Such coordinate representation was reported by Buneo et al. (2002), who found that neurons in the posterior parietal cortex (the parietal area projecting to the PMd) are coding target locations with respect to the hand, as well as the eye.

On the other hand, our findings demonstrating hand-side selectivity emphasize one additional aspect of representation in the PMv, namely the spatial coding of the PMv centered on a specific point in the object to be moved. PMv neurons have been reported to encode the spatial position of a body part (Graziano et al., 1994) or space based on objects that are acted upon (Murata et al., 1997; cf. Galati et al., 2000). This is different to the representation of the relative position of a target within objects that are to be captured (Olson and Gettner, 1995). In our study, the image of the hand was a visual object that was to be moved towards a target in space. Furthermore, hand-side selectivity was unaffected by shifting the initial position of the hand in the video image, which suggests that the neuronal activity in the PMv depended on the spatial relationship of the point of interest referenced both to the moving hand and the target. Therefore, our findings suggest that the PMv determines which part of the body moves in which direction. Namely, the PMv seems to be involved in controlling movement in a coordinate frame that adapts to the reference frame of the object to be moved. PMv neurons appeared to be relatively more involved in specifying a specific point of interest in the moving object than PMd neurons (Table 3), though direct comparison of both areas was not possible. Although our results were obtained under artificial conditions in which hand movements depended crucially on the visual image of the motion of the hand, such behavioral conditions are often applicable to behavioral conditions in everyday life, such as operating a computer mouse or other tools that are used to point to a target that is presented on a video display unit. One weakness of this study is that our findings are based on one animal. Although the sample size of neurons was large enough to identify properties of PMv cells reported, our findings need ideally to be confirmed in a different animal. Thus, interpretation of the present data should proceed with caution.

Based on the present findings, we wish to propose a view that PMv is representing the action of a controlled object in visual coordinates, as a part of a general concept of higher-order representation in cortical motor areas. Such a concept of representation at an abstract level has been proposed in studies reporting cognitive control of spatiotemporal motor variables (Georgopoulos, 2002), in multilevel visuomotor transformations (Caminiti et al., 1998; Buneo et al., 2002; Battaglia-Mayer et al., 2003), and in effector-independent representations of action (Cisek et al. 2003; Fujii et al. 2002).

We thank M. Kurama and Y. Takahashi for technical assistance. This work was supported by CREST, JST, RR2002 and the Ministry of Education, Culture, Sports, Science and Technology, Japan. This work was supported by the Cooperation Research Program of Primate Research Institute, Kyoto University.

References

Battaglia-Mayer A, Caminiti R, Lacquaniti F, Zago M (
2003
) Multiple levels of representation of reaching in the parieto-frontal network.
Cereb Cortex
 
13
:
1009
–1022.
Buneo CA, Jarvis MR, Batista AP, Andersen RA (
2002
) Direct visuomotor transformations for reaching.
Nature
 
416
:
632
–636.
Caminiti R, Ferraina S, Mayer AB (
1998
) Visuomotor transformations: early cortical mechanisms of reaching.
Curr Opin Neurobiol
 
8
:
753
–761.
Cisek P, Crammond DJ, Kalaska JF (
2003
) Neural activity in primary motor and dorsal premotor cortex in reaching tasks with the contralateral versus ipsilateral arm.
Neurophysiology
 
89
:
922
–942.
Dum RP, Strick PL (
1991
) The origin of corticospinal projections from the premotor areas in the frontal lobe.
J Neurosci.
 
11
:
667
–689.
Fogassi L, Gallese V, Fadiga L, Luppino G, Matelli M, Rizzolatti G (
1996
) Coding of peripersonal space in inferior premotor cortex (area F4).
J Neurophysiol
 
76
:
141
–157.
Fogassi L, Gallese V, Buccino G, Craighero L, Fadiga L, Rizzolatti G (
2001
) Cortical mechanism for the visual guidance of hand grasping movements in the monkey: a reversible inactivation study.
Brain
 
124
:
571
–586.
Fujii N, Mushiake H, Tanji T (
2002
) Distribution of eye- and arm-movement related neuronal activity in the supplementary eye field, and in the supplementary and presupplementary motor areas of monkeys.
J Neurophysiol
 
87
:
2158
–2166.
Galati G, Lobel E, Vallar G, Berthoz A, Pizzamiglio L, Bihan D (
2000
) The neural basis of egocentric and allocentric coding of space in humans: a functional magnetic resonance study.
Exp Brain Res
 
133
:
156
–164.
Gentilucci M, Fogassi L, Luppino G, Matelli M, Camarda R, Rizzolatti G (
1988
) Functional organization of inferior area 6 in the macaque monkey.I. Somatotopy and the control of proximal movements.
Exp Brain Res
 
71
:
475
–490.
Georgopoulos AP (
2002
) Cognitive motor control: spatial and temporal spects.
Curr Opin Neurobiol
 
12
:
678
–683.
Godschalk M, Lemon RN, Kuypers HGJM, Van Der Steen J (
1985
) The involvement of monkey premotor cortex neurones in preparation of visually cued arm movement.
Behav Brain Res
 
18
:
143
–158.
Graziano MS, Yap GS, Gross CG (
1994
) Coding of visual space by premotor neurons.
Science
 
266
:
1054
–1057.
Hoshi E, Tanji J (
2002
) Contrasting neuronal activity in the dorsal and ventral premotor areas during preparation to reach.
J Neurophysiol
 
87
:
1123
–1128.
Iriki A, Tanaka S, Obayashi Y, Iwamura Y (
2001
) Self-images in the video monitor coded by monkey intraparietal neurons.
Neurosci Res
 
40
:
163
–173.
Kakei S, Hoffman DS, Strick PL (
2001
) Direction of action is represented in the ventral premotor cortex.
Nat Neurosci
 
4
:
1020
–1025.
Kurata K (
1991
) Corticocortical inputs to the dorsal and ventral aspects of the premotor cortex of macaque monkeys.
Neurosci Res
 
12
:
263
–280.
Kurata K, Hoshi E (
1999
) Reacquisition deficits in prism adaptation after muscimol microinjection into the ventral premotor cortex of monkeys.
J Neurophysiol
 
81
:
1927
–1938.
Kurata K, Hoshi E (
2002
) Movement-related neuronal activity reflecting the transformation of coordinates in the ventral premotor cortex of monkeys.
J Neurophysiol
 
88
:
3118
–3132.
Lacquaniti F, Caminiti R (
1998
) Visuo-motor transformations for arm reaching.
Eur J Neurosci
 
10
:
195
–203.
Luppino G, Murata A, Govoni P, Matelli M (
1999
) Largely segregated parietofrontal connections linking rostral intraparietal cortex (areas AIP and VIP) and the ventral premotor cortex (areas F5 and F4).
Exp Brain Res
 
128
:
181
–187.
Matelli M, Camarda R, Glickstein M, Rizzolatti G (
1986
) Afferent and efferent projections of the inferior area 6 in the macaque monkey.
J Comp Neurol
 
251
:
281
–298.
Moll L, Kuypers HG (
1977
) Premotor cortical ablations in monkeys: contralateral changes in visually guided reaching behavior.
Science
 
198
:
317
–319.
Murata A, Fadiga L, Fogassi L, Gallese V, Raos V, Rizzolatti G (
1997
) Object representation in the ventral premotor cortex (area F5) of the monkey.
J Neurophysiol
 
78
:
2226
–2230.
Murata A, Gallese V, Luppino G, Kaseda M, Sakata H (
2000
) Selectivity for the shape, size, and orientation of objects for grasping in neurons of monkey parietal area AIP.
J Neurophysiol
 
83
:
2580
–2601.
Mushiake H, Tanatsugu Y, Tanji J (
1997
) Neuronal activity in the ventral part of premotor cortex during target-reach movement is modulated by direction of gaze.
J Neurophysiol
 
78
:
567
–571.
Ochiai T, Mushiake H, Tanji J (
2002
) Effects of image motion in the dorsal premotor cortex during planning of an arm movement.
J Neurophysiol
 
88
:
2167
–2171.
Olson CR, Gettner SN (
1995
) Object-centered direction selectivity in the macaque supplementary eye field.
Science
 
269
:
985
–988.
Rizzolatti G, Camarda R, Fogassi L, Gentilucci M, Luppino G, Matelli M (
1988
) Functional organization of inferior are 6 in the macaque monkey.
Exp Brain Res
 
71
:
491
–507.
Shen L, Alexander G (
1997
) Preferentiala representation of instructed target location versus limb trajectory in dorsal premotor area.
J Neurophysiol
 
77
:
1195
–1212.
Schwartz AB, Moran DW, Reina GA (
2004
) Differential representation of perception and action in the frontal cortex.
Science
 
303
:
380
–383.
Weinrich M, Wise SP, Mauritz KH (
1984
) A neurophysiological study of the premotor cortex in the rhesus monkey.
Brain
 
107
:
385
–414.
Wise SP (
1985
) The primate premotor cortex: past, present, and preparatory.
Annu Rev Neurosci
 
8
:
1
–19.
Kurata K (
1993
) Premotor cortex of monkeys: set- and movement-related activity reflecting amplitude and direction of wrist movements.
J Neurophysiol
 
69
:
187
–200.
Mushiake H, Inase M, and Tanji J. (
1991
) Neuronal activity in the primate premotor, supplementary and precentral motor cortex during visually guided and internally determined sequential movements.
J Neurophysiol
 
66
:
705
–718.

Author notes

1Department of Physiology, Tohoku University school of Medicine, Japan, 2Miyagi Organization for Industry Promotion, Sendai, Japan and 3The Core Research for Evolutional Science and Technology Program, Japan