Interactions within and between parallel parietal-frontal networks involved in complex motor behaviors in prosimian galagos and a squirrel monkey

Abstract

Long-train intracortical microstimulation (ICMS) of motor (M1) and posterior parietal cortices (PPC) in primates reveals cortical domains for different ethologically relevant behaviors. How functional domains interact with each other in producing motor behaviors is not known. In this study, we tested our hypothesis that matching domains interact to produce a specific complex movement, whereas connections between nonmatching domains are involved in suppression of conflicting motor outputs to prevent competing movements. In anesthetized galagos, we used 500-ms trains of ICMS to evoke complex movements from a functional domain in M1 or PPC while simultaneously stimulating another mismatched or matched domain. We considered movements of different and similar directions evoked from chosen cortical sites distant or close to each other. Their trajectories and speeds were analyzed and compared with those evoked by simultaneous stimulation. Stimulation of two sites evoking same or complementary movements produced a similar but more pronounced movement or a combined movement, respectively. Stimulation of two sites representing movements of different directions resulted in partial or total suppression of one of these movements. Thus interactions between domains in M1 and PPC were additive when they were functionally matched across fields or antagonistic between functionally conflicting domains, especially in PPC, suggesting that mismatched domains are involved in mutual suppression. Simultaneous stimulation of unrelated domains (forelimb and face) produced both movements independently. Movements produced by the simultaneous stimulation of sites in domains of two cerebral hemispheres were largely independent, but some interactions were observed.

NEW & NOTEWORTHY Long trains of electrical pulses applied simultaneously to two sites in motor cortical areas (M1, PPC) have shown that interactions of functionally matched domains (evoking similar movements) within these areas were additive to produce a specific complex movement. Interactions between functionally mismatched domains (evoking different movements) were mostly antagonistic, suggesting their involvement in mutual suppression of conflicting motor outputs to prevent competing movements. Simultaneous stimulation of unrelated domains (forelimb and face) produced both movements independently.

INTRODUCTION

In the present study, we used electrical stimulation of specific locations in parietal and frontal cortex of monkeys to address the basic question of how different motor behaviors are generated. There is a long history of using electrical stimulation of premotor (PMC), motor (M1), and posterior parietal (PPC) regions of cortex to reveal their roles in motor behavior. Early studies of motor cortex revealed a crude, overall map of mainly contralateral body part movements that progressed from foot to hand to face in a mediolateral direction across cortex (Leyton and Sherrington 1917; Penfield and Boldrey 1937; Woolsey et al. 1952) that was based on current levels that were close to threshold and consisted of a brief series of pulses. Leyton and Sherrington (1917) referred to the resulting maps as one of first movements, and longer periods of stimulation resulted in more prolonged movements that involved more body parts. More recently, Graziano and colleagues used longer trains of electrical stimulation of motor and premotor cortex in macaque monkeys and found that different complex movements could be evoked from different parts of motor cortex when stimulated with a half-second series of electrical pulses (see Graziano and Aflalo 2007 for review). However, even the near-threshold stimulations of motor cortex revealed a patchwork or mosaic of groups of columns of first movements of different body parts in motor cortex as possible components of complex movements (e.g., Gould et al. 1986; see Schieber 2001 for review). There also has been a long-standing interest in the motor functions of posterior parietal cortex (Mountcastle et al. 1975), and a dorsal stream of visual processing (Goodale and Milner 1992; Ungerleider and Mishkin 1982) was proposed to provide posterior parietal cortex the visual information needed to guide arm and hand movements for grasping, reaching, and manipulation of objects (Johnson et al. 1996; Matelli et al. 1998). Studies of posterior parietal cortex provided evidence for subregions of cortex near the intraparietal sulcus that are specialized for contributing to different actions such as reaching (Batista et al. 1999; Johnson et al. 1996; Kalaska et al. 1983), looking (Thier and Andersen 1998), grasping (Sakata et al. 1999), or defensive movements of the head and arm (Cooke et al. 2003). Much of the evidence for these movement-associated centers came from relating the recorded responses of neurons to behaviors, but the “defensive” region in ventral intraparietal area (VIP) was best defined by the forelimb defensive (protective) movements that were evoked with longer trains of electrical pulses (Cooke et al. 2003). These movement-related regions of PPC are also known for projecting to parts of premotor and motor cortex, thus giving rise to the popular concept of parieto-frontal processing streams for different complex movements in monkeys and humans (see Caminiti et al. 2015).

Our interest in the organization of frontal-parietal networks was provoked by the early advances of Graziano et al. (2002a, 2002b). It seemed to us that motor cortex was functioning less as a single area and more as a collection of small regions or domains that mediated different behaviors. Our studies focused on studying parietal-frontal networks in prosimian galagos and New World squirrel and owl monkeys, because these primates have more of the frontal and parietal cortex on the surface for the access of visually placed microelectrodes, injections of tracers, and optical imaging. In brief, we found that electrical stimulation in both awake and anesthetized primates evokes complex movements in three regions, premotor cortex, motor cortex, and posterior parietal cortex. Sites for the same or similar movements are grouped, forming cortical domains, and each region has a matching set of eight or more functionally distinct domains that are similar in both galagos and New World monkeys. They also appear to exist in macaque monkeys in a slightly rotated pattern so that the reaching region is posterior (and not medial) to the grasping region. Electrical stimulation of PPC domains selectively activates functionally matched premotor and motor domains (Stepniewska et al. 2011), and feedforward connections from PPC to premotor and motor regions are focused on matching domains (Stepniewska et al. 2009b). Stimulation of PPC domains does not activate other domains, although they are interconnected, suggesting that these longer intrinsic connections activate inhibitory neurons. These and other related results are described and reviewed elsewhere (Kaas et al. 2012, 2013), and a model of the interactions of domains have been presented (Kaas and Stepniewska 2016). We proposed that the different functional domains in each region interact mainly via mutual suppression. Thus the domains at each level are in competition with each other for mediating one of several possible behavioral outcomes. The dominance of a particular PPC domain over others would strongly bias functionally matched premotor and motor domains, because they also compete with each other. We propose that the outcome of the PPC competition is largely based on sensory information, but premotor and motor domains have access to other sources of information, and thus the final outcome may change. Lesioning or inactivating motor cortex domains selectively inactivates matching PPC domains (Cooke et al. 2015; Stepniewska et al. 2014a; I. Stepniewska, R. M. Friedman, A. W. Roe, and J. H. Kaas, unpublished observations).

To evaluate our model, and to further determine the properties of the proposed frontal-parietal networks, we decided to electrically stimulate two sites in two domains at once. We hypothesized that if we stimulated two functionally matched domains, such as a reach domain in PPC and a reach domain in M1, the reach movement would be facilitated so that the action would be evoked at lower levels of current, and the movement would be more pronounced or faster. Alternatively, if two mismatched domains in the same or different regions were stimulated, such as a reach domain and a hand-to-mouth domain, the two domains would interfere with each other, producing little or no purposeful movement, or perhaps one movement overriding the other. As a third possibility, two stimulation sites with similar movements, such as reach upward and reach forward, could interact to produce an in-between movement of up and out. Also, two different movements, such as a reach and a grasp, might be serially combined so that the reach is followed by a grasp. Finally, similar or different movements of contralateral body parts might be simultaneously evoked from pairs of domains in different cerebral hemispheres with little or no apparent interaction, although interhemispheric connections exist. Because outcomes may be influenced by differences in the current levels or the timing of the onsets of stimulations of the two cortical sites, there is obviously much that can be done. In this article we present the completed results of a subset of many possible ways that dual stimulation can be used to reveal the network properties of the frontal-parietal networks. An abstract of some of the results was presented earlier (Stepniewska et al. 2014b).

MATERIALS AND METHODS

The data reported were obtained from six adult galagos of both sexes and one female squirrel monkey. The study was approved by the local ethics committee and conformed to the procedures outlined in the Guide for the Care and Use of Laboratory Animals.

Surgery

Animals were anesthetized to surgical level with isoflurane. They were then placed in a stereotactic head holder with the body resting on a raised surface, and arm and legs free to move. Although the head was prevented from moving, movements of the eyes, ears, and jaw were possible. Under sterile conditions, the skull was opened, the dura was retracted, and a digital photograph of the exposed cortex was taken. The exposed cortex was protected from desiccation by a thin layer of silicone oil. Isoflurane anesthesia was replaced with ketamine (30–50 mg·kg⁻¹·h⁻¹ iv, constant-rate infusion) anesthesia. The frontal and posterior parietal regions were explored with stimulating microelectrodes, and sites of microstimulation (ICMS) were marked on a high-resolution photograph of the exposed cortex. Our previous studies indicated that evoked forelimb reaching movements are to a constant location in space regardless of initial hand location, but after each movement, the hand was replaced near the body.

Microstimulation Procedure

Stimulation was applied by a Master 8 stimulator (A.M.P.I.) with two biphasic stimulus isolators (Bak Electronics Inc.). Current was delivered with two low-impedance tungsten microelectrodes (MicroProbe, Inc.) inserted perpendicular to the cortical surface to a depth of 1.5–1.8 mm, a level that was optimal for eliciting motor responses. Trains of electrical pulses 500 ms in duration, long enough to evoke complex, coordinated movements similar to those of natural behavior, were delivered to M1 or/and PPC. The duration of single pulses was 0.4 ms, and the pulse rate was set at 300 Hz. At the start of each stimulation session, we first stimulated many sites in the area of interest to determine sites that would be of further interest to pair in joint stimulation. Thus cortical sites within M1 and PPC that evoked distinctly different or similar movements were identified and chosen for pairing (Fig. 1). After each of two sites were stimulated (at least 3 times) and thresholds were found, sites were stimulated simultaneously (also 3 times). For most of the pairs, after simultaneous stimulation, each of the sites were again stimulated separately (in reverse order, 3 times). Thus the general pattern was to stimulate one site, then the other, then both sites together, and again one and then the other site. Currents as low as 10 µA evoked visible coordinated movements from M1, and threshold values were typically between 10 and 40 µA. Stimulus currents never exceeded 120 µA. For PPC, currents used were higher, usually 100–300 µA. Paired sites were simultaneously stimulated at various stimulus intensities (at the current thresholds and above) to determine the nature of their interaction.

Fig. 1.

Three of the types of paired sites that were costimulated. A: locations of functional motor domains in frontal and posterior parietal cortical regions of a galago brain are marked with matching colors. Motor cortex of galagos is bounded by the frontal anterior (FSa) and posterior (FSp) sulci; posterior parietal cortex (PPC) is the region around the horizontally oriented intraparietal sulcus (IPS). Other cortical areas and sulci are shown for reference. Black box enclosing primary motor area (M1), rostral PPC, and other fields indicates the region shown in B–D. B–D: examples of pairs of sites that were costimulated in M1, PPC, and M1+PPC are indicated by the location of electrodes (black arrowheads). The fourth type, costimulating a site in each hemisphere, is not illustrated (see text). 1-2, 3a, and S1, somatosensory cortical areas; agg, aggressive; fl, forelimb; hl, hindlimb; LS, lateral sulcus; MT, V2, and V1, visual cortical areas; PMD and PMV, dorsal and ventral premotor areas; PPCr and PPCc, rostral and caudal PPC; tr, trunk. Domains are based on Stepniewska et al. (2005, 2014a).

Analysis of Evoked Movements

Movements evoked by microstimulation were observed by two or three members of the research team and classified according to type (e.g., hand to mouth, reaching, defensive, etc.). Movements evoked by paired stimulation were further categorized accordingly to type of the paired stimulation movement as combined with features of both single stimulation movements, suppressed so that features of one single stimulation movement did not appear, alternated so that both movements occurred in sequence, facilitated so that movement was faster or farther, or those with no interaction. All movements were recorded on videotape at 30 frames/s. For each stimulation site or pairs of sites, movement sequences (mostly forelimb) were synchronized with the onset of pulse trains signaled by a light-emitting diode (LED) and captured by a digital video camera (Sony; model no. DCR-HC65) mounted directly in front of the animal. Thus the start of the stimulation train could be determined to the nearest video frame. Videotapes were reviewed to verify the type of movements evoked as well as their complexity and time courses. Evoked movements were distinguished from the occasional spontaneous movements, which sometimes occur in ketamine-anesthetized animals, by the consistency of the latency and the movement sequence of the evoked behaviors. In most cases our visual observations of movements were confirmed by frame-by-frame videotape measurements of hand trajectories in space, and speeds of movements were calculated. Movement sequences were drawn from video frames by tracking the anatomical element of the hand that was visible through all phases of movement (usually the tips of digits 2–4 or interphalangeal joints of these digits). To calculate the speed of movements, first we divided each trajectory into sections that corresponded to a distance between start and endpoints of the movement seen in one frame; second, we divided the length of each section by frame duration, which equals 33 ms. We made the same calculation for all sections within the trajectory and prepared the graphs comparing the speeds of the two single site stimulation movements with movement evoked by dual stimulation. Examples of hand trajectories and movements speeds are illustrated for two cases, G14-10 and G15-03 (see Figs. 3–5, 7, 8, and 10–12).

Fig. 3.

Dual stimulation of primary motor area (M1) sites (left hemisphere) evoking different complex movements (right forelimb). Examples are from cases G15-03 and G14-10. The trajectories of movements were evoked from each of the 2 sites stimulated separately (green and blue lines) and concurrently (dashed red line). Arrowheads mark the starting locations of the galago’s right hand. The trajectories proceed upward-downward, away from body, and toward body. The trajectories do not reflect the extents of the forward and backward movements, because the video recordings were done with the camera in front of the animal. Numbers refer to paired intracortical microstimulation (ICMS) sites shown in Fig. 2. A and B show examples of movement summation, and C and D show examples of movement competition, resulting in total (C) or partial (D) dominance of one of the movements. Spacing of the dots reflects the frame time (33 ms). stim., Stimulation amplitudes during paired ICMS; thresh., threshold stimulation amplitude (µA) during single-site characterization.

Fig. 5.

Dual stimulation of primary motor area (M1) sites evoking similar movements leads to the summation/potentiation of movements. A and B: both pairs show that dual stimulation led to a movement (dashed line) similar to pretested movements, but with a larger magnitude (longer trajectory). The trajectory for the costimulation also follows more closely the trajectory of one of the two movements. Spacing of the dots reflects the frame time (33 ms). stim., Stimulation amplitudes during paired ICMS; thresh., threshold stimulation amplitude (µA) during single-site characterization. C and D: speed profiles of movements presented in A and B, respectively, as a function of time during stimulation. Dual stimulation led to faster movements. Examples are from case G14-10. Speed was measured in 33-ms increments. The profile line for the costimulation is dashed. Arrowhead marks start of stimulation.

Fig. 7.

Dual stimulation of posterior parietal cortex (PPC) sites evoking different (A–D) and similar (E) complex forelimb movements. Examples are from cases 14-10 and 15-03. Trajectories of movements are those evoked from each of the 2 sites stimulated separately (green and blue lines) and concurrently (red dashed line). Arrowheads mark the starting locations of the galago’s right hand. Numbers refer to paired intracortical microstimulation (ICMS) sites shown in Fig. 6. A–D show examples of movement competition, resulting in dominance of one of the movements and obstruction of the other movement, and E shows an example of movement facilitation/potentiation. Scale bar, 10 mm. Spacing of the dots reflects the frame time (33 ms). stim., Stimulation amplitudes during paired ICMS; thresh., threshold stimulation amplitude (µA) during single-site characterization.

Fig. 8.

Speed profiles of movements evoked by dual stimulation of posterior parietal cortex (PPC) sites evoking different movements (A) and same movements (B) in case G14-10. Speed was measured in 33-ms increments. The profile line for the costimulation is dashed. Arrowhead marks start of stimulation.

Fig. 10.

Dual stimulation of primary motor area (M1) and posterior parietal cortex (PPC) sites in G15-03 and G14-10. Trajectories of movements are those evoked from each of the 2 sites when stimulated separately (green and blue lines) and concurrently (dashed red line). Arrowheads mark the starting locations of the galago’s right hand. Numbers refer to paired intracortical microstimulation (ICMS) sites shown in Fig. 9. A–C show examples of movements that combined during dual stimulation. D–G show examples of movements that competed during dual stimulation, resulting in the obstruction of one movement by the other. Spacing of the dots reflects the frame time (33 ms). stim., Stimulation amplitudes during paired ICMS; thresh., threshold stimulation amplitude (µA) during single-site characterization.

Fig. 12.

Trajectories (A) and speed profiles (B) of movements evoked by single and dual stimulation of paired sites in primary motor area (M1) and posterior parietal cortex (PPC) representing same/similar movements in case G14-10. Dual stimulation led to faster movements. In A, trajectories of movements are those evoked from each of the 2 sites when stimulated separately (green and blue lines) and concurrently (dashed red line). Arrowheads mark the starting locations of the galago’s right hand. Spacing of the dots reflects the frame time (33 ms). stim., Stimulation amplitudes during paired intracortical microstimulation ; thresh., threshold stimulation amplitude (µA) during single-site characterization. In B, speed was measured in 33-ms increments. The profile line for the costimulation is dashed. Arrowhead marks start of stimulation.

Perfusion and Histological Processing

At the end of each experimental procedure, animals were killed with an overdose of pentobarbital sodium (Euthasol) and, when areflexive, perfused through the heart with buffered physiological saline followed by 2% paraformaldehyde, and then 2% paraformaldehyde with 10% sucrose. The brain was removed, and the cortex was separated from the brain stem, unfolded, manually flattened between two glass plates (see e.g., Stepniewska et al. 2003), and submerged in 30% sucrose at 4°C for cryoprotection. One to three days later, cortical tissue was frozen and cut parallel to the surface into two series of 40- to 50-μm-thick sections. Alternate sections were stained for myelinated fibers (Gallyas 1979) or cytochrome oxidase (CO; Wong-Riley 1979) to reveal cortical architecture. CO and myelin preparations allow sensorimotor (M1, 3a, and 3b), and visual (MT, V2, and V1) areas to be recognized histologically, as in our previous studies (Collins et al. 2001; Fang et al. 2005; Preuss et al. 1996; Stepniewska et al. 1993). The locations of movement domains within PPC were reconstructed from features in the processed sections. Thus the architectonic borders in CO- and myelin-stained sections were drawn, superimposed and aligned using blood vessels, lesions, or other landmarks. Microlesions (10-μA direct current), placed to mark functional borders, helped the alignment of ICMS maps with cortical architecture. Processed tissue showed no signs of damage at the sites of stimulation, in contrast to sites of fiduciary lesions.

RESULTS

Results presented in this report are based on the microstimulation and comparison of 115 pairs of cortical sites within frontal and posterior parietal regions (Table 1). The goal was to characterize movements produced by separate and combined activations of cortical sites. Observations were made on 107 pairs of sites of the same hemisphere and on 8 pairs of sites, with one site in the left and one in the right hemisphere (marked with asterisk). In motor cortex (M1), we costimulated 62 pairs, in posterior parietal cortex (PPC), 19 pairs, and in M1 + PPC, 26 pairs. The behavioral effects of separate vs. simultaneous stimulation of cortical sites were determined for all pairs of sites. All evoked limb movements were made against gravity, as would natural movements. The repertoire of forelimb movements evoked by long-duration ICMS included those previously described (Stepniewska et al. 2005) as defense (forelimb retraction), reach, lift, hand to mouth (or to body), and grasp. The responses were repeatable from trial to trial, and over the time needed to characterize the cortical sites individually, stimulate them simultaneously, and retest them shortly after simultaneous stimulation (typically about 10 min). Often a site was paired with several other sites for dual stimulation. However, a small difference in response strength was sometimes noticed when the same site was stimulated a number of times during a short period of its pairing with multiple other sites. A movement that resulted from simultaneous stimulation did not depend on stimulation order; thus stimulating site 1, then site 2, and then both together, or site 2, site 1, and both together, produced the same outcome. After paired stimulation, the stimulation of each site separately always produced the pretesting movement. Our goal in these experiments was not to delimit domains or collect nearby sites where stimulation produced highly similar movements, but to compare movements evoked from pairs of sites when stimulated separately or together.

Table 1.

Summary of experimental cases

		M1 + M1			PPC + PPC		M1 + PPC
Animal No.	No. of Pairs	FL	FL + Face	FL + HL	FL	FL + Face	FL	FL + Face
G13-04 (F)	9	9 (1–2 mm)
G13-09 (M)	8	7 (1.5–2.5 mm)	1 (5 mm)
G13-22 (M)	10	8	2
G13-28 (F)	21	12 (1–4 mm)		4 (2.5–3.5 mm)	4 (1.5–4 mm)	1 (7 mm)
SM13-52 (F)	9 + 8*	5 + 3*		1	1 + 3*	2 + 2*
G14-10 (M)	28	8 (1–2 mm)	1 (2.5 mm)	1 (4 mm)	5 (0.7–5 mm)	1 (5.7 mm)	11 (8–12 mm)	1 (11 mm)
G15-03 (M)	22	3 (1.5–3 mm)			5 (2–4 mm)		13 (5–10.8 mm)	1 (9 mm)
All	108 + 8*	52 + 3*	4	6	15 + 3*	4 + 2*	24	2

Data are numbers of pairs of sites in primary motor area (M1), posterior parietal cortex (PPC), or M1 + PPC that evoked forelimb (FL), face, or hindlimb (HL) movements in male (M) or female (F) galagos (G) and a squirrel monkey (SM). Numbers in parentheses indicate range of distances between stimulated sites.

^*Paired sites in left and right hemispheres.

Movements Evoked from Paired M1 Sites

Because anesthetic levels affect threshold levels for evoked movements, and stimulation of sites in M1 most reliably evoked complex movements over long stimulation sessions, we studied the effects of dual stimulation of sites in M1 in all seven primates. For pairing we mostly chose sites that represented different forelimb movements and some representing same/similar forelimb movements (52 pairs; Table 1). In a few instances, we also paired sites for forelimb movements with sites for face (4 pairs) or with sites for hindlimb movements (6 pairs) to study interactions between sites representing movements of two different body parts (Table 1). Together, 62 pairs of M1 sites were studied. The distances between paired M1 sites ranged between 1 and 4 mm (see Table 1). The locations of these sites on a surface outline of the motor cortex for each studied case are shown in Fig. 2. Most often, sites were stimulated at or near their thresholds, but to see movements better during the video recording, we sometimes similarly increased the current at each individual site and then stimulated both sites together with a higher current. Because our goal was to compare movements evoked by pairs of sites stimulated, both separately and together, we chose sites for comparison based on the types of movements that were evoked and did not extensively map parts of M1 to define domain borders. Thus the relationship of sites to the boundaries of domains in M1 was somewhat uncertain. In addition, the arrangement of domains in M1 is somewhat variable across cases and may vary from those depicted in Fig. 1.

Fig. 2.

Distributions of stimulated sites on surface views of motor cortex (M1) in 5 studied galago (G) cases and 1 squirrel monkey (SM) case. Intracortical microstimulation (ICMS) responsive sites are marked by closed circles and unresponsive sites by open circles. Paired ICMS sites described in text are identified by large closed circles and site numbers. Zones of sites with different movement thresholds are shown in different shades of gray. Thick black lines indicate approximate borders of M1, and thin black lines indicate approximate borders of complex movement domains. Motor cortex is bordered by premotor cortex (PMD) and area 3a. Horizontal scale bars, 1 mm. FSa and FSp, anterior and posterior frontal sulci.

Stimulation of sites evoking different forelimb movements.

Most cortical sites in M1 chosen for dual stimulation were located in different forelimb movement domains (41 pairs; Table 2). During simultaneous stimulation of two sites in M1, we often observed the preservation (at least partial) of characteristics of each movement, and their combination resulted in a novel movement. Thus the movement evoked by simultaneous stimulation of two cortical sites representing different movements was often a blend or combination of the individually evoked movements (65.9%), with varying degrees of involvement of the individual movements. However, when the two stimulated sites evoked contradictory movements (e.g., reaching up and reaching down), only one of the two movement patterns occurred during dual stimulation, and the other one was suppressed (29.2%). Rarely, costimulation of two sites in different domains resulted in alternations of the two movements (4.9%). The occurrences of the different movement patterns evoked by dual stimulation are shown in Table 2.

Table 2.

Summary of movements evoked by costimulation of sites representing different and same forelimb movements

	M1 + M1				PPC + PPC				M1 + PPC
	Different			Same	Different			Same	Different			Same
Animal No.	Comb.	Suppr.	Alter.	Facilit.	Comb.	Suppr.	Alter.	Facilit.	Comb.	Suppr.	Alter.	Facilit.
G13-04	6	3
G13-09	1	3	2	1
G13-22	3	1		4
G13-28	5	4		3	1		2	1
SM13-52	4	1				1
G14-10	6			2		4		1	3	7		1
G15-03	2			1		5			4	6		3
All	27 (65.9%)	12 (29.2%)	2 (4.9%)		1 (7.7%)	10 (76.9%)	2 (15.4%)		7 (35%)	13 (65%)
	41 (79%)			11 (21%)	13 (86.6%)			2 (13.3%)	20 (83.3%)			4 (16.7%)
	52				15				24

Data are numbers of pairs of sites in primary motor area (M1), posterior parietal cortex (PPC), or M1 + PPC that evoked different or same/similar movements in galagos (G) and a squirrel monkey (SM). Costimulation of sites representing different movements evoked 3 categories of movements: combined (Comb.), suppressed (Suppr.), and alternated (Alter.). Costimulation of sites for same/similar movements resulted in movement facilitation (Facilit.). Numbers in parentheses indicate percentages of pairs that evoked each category of different and same movements with costimulation of M1 + M1, PPC + PPC, and M1 + PPC. For more details, see text.

combined (blended) movements.

Good examples of combined movements come from cases G15-03 and G14-10. In both cases, stimulated sites in M1 were placed in different domains that were about 2 mm distant from each other. Videotaped movements were analyzed, and trajectories of the movements and their speeds were calculated (see Figs. 3 and 4). In case G15-03, costimulation of sites that evoked a lateral and upward reach (site 40) and a hand retraction (site 42) at their thresholds (10 µA) resulted in the arm being laterally lifted and retracted (Fig. 3A). A similar blending of movements was observed during dual stimulation of sites for a reach upward (site 59) and a lateral defense/avoidance movement (site 61) in case G14-10 (Fig. 3B). The consequence of dual stimulation resulted in a hand that was lifted and moved laterally. Thus, in both cases, dual stimulation evoked a movement that was clearly a blend of the movements evoked by individual stimulation of each site. In both cases, the direction and amplitude of the combined movement was nearly the average of those of the individual movements. The speed of the two combined movements was also the average of the two individual movement speeds (Fig. 4A). In the same case G14-10, dual stimulation evoked a blend of movements for the other five pairs of sites. For example, site 16 represented a reach with a grasp followed by hand-to-mouth movement, and site 58 represented a forelimb retraction. Dual stimulation produced a blended movement that involved reach with grasp and a lift characteristic for site 16, with the addition of a retraction component characteristic for site 58.

Fig. 4.

Speed profiles of movements evoked by dual stimulation of primary motor area (M1) sites. A–C: results of individual site stimulation compared with results of stimulation of the pair of sites together. Hand speed is shown as a function of time during stimulation of 2 different domains in case G14-10. Speed was measured in 33-ms increments. The profile line for the costimulation is dashed. Arrowhead marks start of stimulation.

Paired stimulation of different kinds of reaches (e.g., upward, sideward, or to body) in case G13-28 led to a blend of the individual reach movements, as well (not illustrated). Also, costimulation of sites evoking complementary movements such as lift (site 18) and grasp (site 15) at their thresholds in case 13-04 produced a blended movement with the hand grasping and lifting. For all of these described sites, the stimulated pairs were situated quite close (1–2 mm apart) to each other. However, a blend of movements was also seen during stimulation of more distant sites. Thus stimulation of sites 4 mm apart representing a reach-upward movement (site 5) and a limb-retraction movement (site 11) in case G13-28 produced a blended or combined movement, with the hand lifted while the upper arm still retracted.

These results provide evidence for blending of movements evoked by stimulation of two M1 sites located in different forelimb domains. The results were from pairs of sites that were stimulated at their thresholds or slightly above their thresholds. When the current for one of the stimulated sites, but not the other, was increased during simultaneous stimulation, we observed a stronger expression of the movement with the increased current. Similar results were observed when two sites of different thresholds were stimulated with the same current. In such cases, during costimulation of both sites with current higher than either threshold, the movement with the lower threshold increased, at the cost of the movement of the higher threshold, which decreased in magnitude. Repeating these stimulations showed convincingly that dual stimulation of these sites involves some mixing of the two movements, with a dominance of the lower threshold movement. Such observations were made in three studied animals (G13-28, SM13-52, and G13-04).

All these examples show that movements evoked by simultaneous stimulation of two M1 sites for different movements tend to include some aspects of the movements of both sites in the paired stimulation movement. Thus movements blend into a third movement, combining what appear to be the most salient aspects of each individual movement. However, when current was increased for one of these sites (or one of these sites had a lower threshold), the movement represented by this site was more pronounced in such a blend. Thus aspects of both movements were present, but one of the movement types was stronger than another. Similar phenomena have been described previously by others when two sites were simultaneously stimulated in cat M1 (Ethier et al. 2006), monkey frontal eye field (Robinson and Fuchs 1969), or superior colliculus (Robinson 1972).

movement suppression.

Less often during costimulation of two M1 forelimb sites, we observed a suppressive effect of one site over the other (29.2%; Table 2). When the two stimulated sites evoked more or less contradictory movements, only one of the two movement patterns occurred during dual stimulation. A strong suppression effect was observed in case G14-10 when we paired two sites evoking contradictory movements, reaching up (evoked from site 59) and reaching down (evoked from site 60) (Fig. 3C). These two sites were stimulated respectively with currents three and two times higher than their thresholds. As shown in Fig. 3C, the reach movement evoked from site 59 clearly dominated the movement evoked from site 60. The trajectory of the paired stimulation movement closely followed the trajectory of the movement evoked from site 59. This trajectory was also longer than the trajectory of movement evoked from site 59, suggesting the disinhibition of site 59, but speed of the movement remained the same (Fig. 4B). One may think that suppression of one site of the pair by the other could be due to the higher stimulation current applied to the “winning site,” given that this site (59) was stimulated with current three times higher than threshold, relative to two times higher than threshold for another site (60). Indeed, our other experiments with numerous pairs of sites have shown that increasing the current of stimulation for one site may more drastically evoke total or near total suppression of the other site. However, a suppressive effect was also observed when both sites were stimulated at their thresholds or proportionally over their threshold. For example, in case G13-28, costimulation of two M1 sites (Fig. 2), for a grasp and lift (site 30) and a reach forward (site 31) at their threshold, resulted in a grasp and lift movement. Thus the grasp-with-lift site suppressed the other reach site, resulting in complete loss of the reach movement. Although “losing” site 31 was quite rostral and could be located in dorsal PMC, a similar effect was observed for pairs of other sites, when both sites were clearly in M1.

More complex suppressive effect of dual stimulation was observed in case G14-10, when we paired a site representing upward reaching movement (59) with a site for arm retraction and grasp (62). Simultaneous stimulation evoked only a movement that retained the arm retraction coded by site 62, but without the grasp (Fig. 3D). Thus, although the movement for site 62 seemed to win the competition, there was some evident inhibitory effect from the other site, resulting in loss of the grasp component. This inhibitory effect was also expressed as a decrease in the speed of the prevailing movement (Fig. 4C). Interestingly, during pairing, site 62 (the “winning” site) was stimulated with current two times higher than threshold, and site 59 (the “losing” site) was stimulated with current three times higher than its threshold. Thus suppression is not necessarily dependent on the increase of stimulating current.

All the above-described sites in pairs that suppressed each other were positioned quite close together (2.7 mm or less apart). When stimulated sites were separated farther from each other, we could still observe interference. However, this resulted only in a partial dominance of one movement over the other. For example, in squirrel monkey SM13-52, where we paired two sites about 4 mm apart, one for a slightly lateral forelimb reach up (site 97) and another site for hand-to-mouth movement (site 31) (Fig. 2), simultaneous stimulation (at their thresholds, which happened to be the same, 10 μA) evoked a movement that retained more of the hand-to-mouth movement (as site 31) and less of the lateral upward reach movement (as site 97).

In summary, during dual stimulation of two different forelimb domains, one domain might suppress the other, leading to generation of a movement that is similar to the movement evoked from one of the stimulated domains, either without the other movement component or with this component being much weaker than during pretesting.

alternated movements.

In only a few instances (4.9%), stimulation of sites in two different M1 domains at once resulted in an alternation of movements. For example, for two sites in case G13-09 (not illustrated), the movement represented at one site (e.g., reach up) was followed by the movement represented at the other site (e.g., retraction). This alternation of movements was clear and repeatable throughout the trials.

Stimulation of sites evoking the same/similar kind of forelimb movement.

For 11 pairs (21%; Table 2), stimulated sites were located in the same forelimb movement domain in M1, thus evoking the same or very similar forelimb movements during pretesting (e.g., arm lift). Because of the small size of movement domains, sites chosen for stimulation were usually situated close to each other (<2 mm apart). ICMS applied simultaneously to both sites most often caused a movement of a same/similar direction, but of a larger magnitude. For example, in case G14-10 (Fig. 5A), we stimulated two sites (66 and 67), each evoking upward reach movement, with the same threshold of 30 µA. These sites were just over 2 mm apart (Fig. 2). To see movements better during the video recording, we doubled the current, and each individual site and then both sites together were stimulated with current of 60 µA. Under this condition, simultaneous stimulation of sites 66 and 67 evoked the reach movement, similar to movements evoked from either site, but of larger magnitude. Thus the movement trajectory was longer, and the speed of this movement was much higher than the speed of individual movements (see Fig. 5, A and C). When we switched back to stimulation of sites individually, the evoked movements returned to their normal pretesting state (lower magnitude, lower speed). Similar movement facilitation was observed when we stimulated two other closely positioned sites (59 and 64) in the same case G14-10. Trajectories of evoked movements are illustrated in Fig. 5B. Again, both of these sites represented upward reaching movement, but one of the movements was directed slightly laterally. As a consequence of dual stimulation, we observed the more pronounced reach. The speed of this movement was also faster than the speeds of individual movements (Fig. 5D). Similar movement facilitation/enhancement was observed for four other pairs of sites located in reach domains in other cases (G13-22 and G13-28). For all these pairs, stimulation at thresholds generated a reach movement of higher magnitude. Thus stimulating two sites at once that separately evoke similar movements potentiates each other to evoke a movement that is greater (with longer trajectory) and faster (of higher speed) even when the two sites are about 2 mm apart.

Most stimulation-evoked M1 forelimb movements were initiated during the first two video frames, thus within 66 ms after the electric stimulus onset. This movement initiation profile was generally conserved during dual stimulation. In case G14-10, where movement speeds were calculated, only in one instance (sites 59+60; Fig. 4B) was a delay detected, and it amounted to a single video frame. Interestingly, this was the case when paired movements were contradictory (lift and reach down). Usually, all movements evoked during individual or double stimulation of M1 sites started at the same time after stimulus onset. When the movement evoked from one site was delayed in relation to the movement evoked from the other site (sites 66+67), the movement evoked during stimulation of both sites at the same time started as first movement (Fig. 5C). Speeds of movements evoked during simultaneous stimulation of sites evoking different movements were either average (e.g., sites 59+61; Fig. 4A), the same (e.g., sites 59+60; Fig. 4B), or similar to the speed of the wining movement (e.g., sites 59+62; Fig. 4C). When sites evoking similar movements were paired, speeds of the new movements were much higher (for sites 66+67; Fig. 5C) or somewhat higher (for sites 59+64; Fig. 5D) than the speed of individually evoked movements.

Pairing sites for forelimb movements with those for face or hindlimb movements.

We stimulated 10 pairs of M1 sites that evoked forelimb movements, together with sites evoking face (4 pairs) or hindlimb movements (6 pairs) (see Table 1). These sites were usually more distant from each other than the pairs of sites in forelimb representation (see above). Most often, costimulation of M1 sites representing movements of different body parts showed a lack of visible interaction between them. However, for some less distant sites, some interactions were noticed.

Simultaneous stimulation of sites 4–5 mm apart that evoked a forelimb reach and a face grimace or tongue movement (cases G13-09 and G13-22) resulted in two movements represented by each individual site, and of the same magnitude as during pretesting. Thus we did not observe a clear interaction between movements evoked from these two sites. Another example of such lack of interaction was observed between sites that evoked forelimb and hindlimb movements (cases G14-10 and SM13-52). In both cases, dual stimulation of such sites resulted in the same forelimb and hindlimb movements, but they were performed simultaneously.

Different results were found for sites that were situated closer to each other (about 2–3 mm apart). During dual stimulation of such sites, we observed an increase of one movement at the cost of another movement. For example, in case G13-28, stimulation of sites for forelimb reach and hindlimb movement at their thresholds caused an increase of the amplitude of hindlimb movement, with dramatic reduction of reach movement. Thus the hindlimb site seemed to suppress the forelimb site, which in turn appeared to disinhibit the hindlimb site, resulting in a larger hindlimb movement. When stimulation of the hindlimb, but not the forelimb, site stopped, the reach movement came back as strong as that during pretesting. The results were similar for three other pairs of sites in the same case, G13-28. During simultaneous stimulation of forelimb and hindlimb sites, forelimb movements were always inhibited, whereas hindlimb responses increased in size.

A somewhat different (more complex) interaction was observed between stimulated sites responsible for movements of different body parts that were less distant from each other, e.g., sites responsible for the hand movement and tongue movement, situated about 2.5 mm apart (case G14-10). Although the hand movement seemed to remain at the same strength during concurrent stimulation, the tongue movement became more pronounced. All these examples suggest that sites in M1 for the movements of different body parts also interact, sometimes with a partial reduction of suppression.

In summary, our analysis of movements evoked from sites in M1 (Figs. 3–5) supports the following conclusions. First, when two stimulated M1 sites represented different movements, the resultant movement most often was a blend of the individually evoked movements, with the trajectory and speed close to the average of the individual movements. Less often, one of the movements suppressed the other one, so the resulting movement from dual stimulation was similar to one of the two movements, with a similar trajectory and speed. Only sporadically did dual stimulation of paired sites for different forelimb movements evoke an alternation of movements. Second, for two sites that produced the same or similar movement, dual stimulation resulted in new movement of the same kind, but more pronounced. Trajectory of the resultant movement was longer and the speed higher than those of individually evoked movements. Third, costimulation of two cortical sites representing movements of two different body parts and separated by 4 mm or more did not reveal any obvious interaction. Thus movements remained the same. If such two sites were closer to each other, some suppression of one site by the other site could take place, with one of the movements being stronger and the other weaker than during pretesting.

Movements Evoked from Paired PPC Sites

To study interactions between functional PPC domains, as for M1, we chose sites that represent different or similar movements of forelimb. Interactions between forelimb sites were studied in 15 pairs of sites, and in 4 other pairs we studied interactions between sites representing movements of forelimb and face (Table 1). The distances between PPC paired sites ranged between 0.7 mm for sites belonging to the same functional domain and 7 mm for sites that belonged to different movement domains, such as reach and face grimace domains. For most paired sites, the distance between them was about 4 mm. As for M1, PPC sites were stimulated at or somewhat over the threshold. The distributions of all stimulated sites are shown on surface outlines of PPC in Fig. 6.

Fig. 6.

Stimulated sites in rostral posterior parietal cortex (PPCr) in 3 galagos (G13-28, G14-10, and G15-03) and 1 squirrel monkey (SM13-52). In galagos, the border between rostral and caudal PPC is marked by a dashed line. In squirrel monkey, PPC is located posterior to central sulcus and immediately anterior to lateral sulcus (LS) and in its depth. Intracortical microstimulation (ICMS) responsive sites are marked by closed circles and unresponsive sites by open circles. Paired ICMS sites described in text are identified by large closed circles and site numbers. Zones of sites with different movement thresholds are shown in different shades of gray. IPS, intraparietal sulcus.

Stimulation of PPC sites evoking different forelimb movements.

Thirteen pairs (86.6%) of PPC sites evoked different forelimb movements when both sites were stimulated (Table 2). As for M1, we observed three different classes of interactions evoked by dual stimulation of PPC different forelimb domains. However, whereas for M1 pairing, the most common response to such stimulation was a blend of movements (65.9%), for PPC pairing, it was domination of one movement over the other (76.9%). For PPC pairing, a blend (7.7%) or alternation of movements (15.4%) was seen only sporadically.

In case G14-10, we paired a site for reach up and forward (site 42) with a site producing hand-to-body movements (site 51) or with a site for hand-to-mouth movements (site 49) (Fig. 6). Simultaneous stimulation of the paired sites resulted in a domination of hand-to-body (Fig. 7A) or hand-to-mouth movements over the reach (not illustrated). In both events, the trajectory of movements evoked by dual stimulation, although a little shorter, followed the trajectory of hand-to-body or to-mouth movements evoked by individual stimulation of sites 51 or 49. Yet, some suppression occurred. As illustrated in Fig. 8A, for the pair of sites 42 and 51, the highest speed of the movement that resulted from dual stimulation (36.4 cm/s) was comparable to the average of speeds of two separately tested movements (site 42, 25.8 cm/s, and site 51, 45.5 cm/s). A similar suppressive effect was observed during simultaneous stimulation of sites 6 and 15 in case G15-03 (Fig. 6) for reach-up and hand-to-body movements, respectively (Fig. 7B). For this pair, site 6 also was completely suppressed, although it was stimulated with proportionately higher current than the paired site (site 15). As for the pair of sites 42 and 51, the new movement followed the hand-to-body movement, although the trajectory again was a bit shorter in this case. Costimulation of two other pairs of PPC sites in case G14-10 (sites 42+46 and 42+47; Fig. 6) that evoked forelimb movements in opposite directions (reaching forward and retraction) also caused clear suppression of one of the movements (reach). This suppression occurred even though the retraction movements evoked from individually stimulating sites 46 or 47 were less extensive than the reach movement evoked from site 42.

A domination of one site over the other also was seen when site 6 (reach up) competed with two other sites, site 13 (hand to body) and site 16 (arm retraction), in case G15-03 (Fig. 6 and Fig. 7, C and D). In both circumstances the reach movement lost to the other movements. The trajectories of winning movements highly overlapped those of the single site stimulation, and they were a little longer compared with the pretesting situation. Similar results were obtained for PPC sites representing lateral reach (site 105) and wrist supination with grasp (site 103) in monkey SM13-52 (Fig. 6) when during dual stimulation we observed an exaggerated (longer) lateral reach, but not supination or grasp (not illustrated). In all mentioned cases, when the stimulation current for the winning site was lowered below its threshold, the movement associated with the other site became visible. Overall, suppressive interactions between PPC domains, similar to those of M1 domains, can be complex, but disruptions of movements evoked from one of the two costimulated domains in PPC were greater than those evoked from M1.

Other effects, such as the alternation or blend of different forelimb movements when two PPC sites were simultaneously stimulated, were rare and found only in one galago (G13-28; Table 2). Stimulation of two pairs of sites that combined the reach movement (site 21) with either hand-to-body (site 23) or hand-to-mouth movements (site 26) evoked sequential or alternated movements, so first we observed one movement (e.g., reach) and then the second movement (e.g., hand to body). This sequential response was very clear and repeatable from trial to trial. Another dual stimulation of a reach site (site 18, threshold 150 µA) and a hand-to-body site (site 20, threshold 200 µA) evoked a blend of both movements, although the reach was less pronounced, whereas the movement of the hand toward the body was more obvious.

Stimulation of sites evoking similar forelimb movements.

In a couple of instances, we stimulated two PPC sites located in the same movement domain that evoked the same or a very similar forelimb movement from each site. In case G14-10, concurrent stimulation of two closely positioned PPC sites (sites 26 and 53, about 1 mm apart), both representing a hand-lift movement, evoked a similar but stronger lift with a longer trajectory (Fig. 7E). The maximal speed of this movement (Fig. 8B) was also much higher (42.4 cm/s) than maximal speeds of the individual movements, which were similar for both stimulated sites (25.8 and 27.3 cm/s). Similarly, in another case (G13-28), stimulation of two slightly more distant sites (sites 18 and 21, 1.25 mm apart), both evoking reach movements, produced a similar reach movement of a larger magnitude and greater top speed (not shown). These results are comparable to results we obtained when stimulating M1 sites within the same movement domain.

Our analysis of the initiation and speed of movements evoked from PPC forelimb domains is quite limited. Two analyzed pairs of movements have shown that movements evoked from PPC were initiated during the first three or four video frames, thus within 99–132 ms after the stimulus onset. As for M1, the movement initiation profile was generally conserved during dual stimulation. However, the latency of the movements evoked by costimulation was not less than that for the separately evoked movements (~100 ms). When two domains evoking different movements were stimulated at once, the speed of resulting movements was averaged (sites 42+51; Fig. 8A). The speed of movements evoked during costimulation of sites in the same domain was higher compared with speeds of individually evoked movements (sites 26+53; Fig. 8B).

Stimulation of sites for forelimb and other body part movements.

In cases where we combined stimulation of sites evoking movements of forelimb and face, we did not observe any interaction between those movements, and both movements were similar to the pretesting condition. For example, stimulation of sites representing a reach (site 21) and a face grimace (site 29), separated by a distance of 7 mm (case G13-28; Fig. 6), evoked both movements at once, similar to movements observed when these sites were stimulated separately. The same result occurred when a reach forward (site 103) was paired with a face grimace (site 106) in case SM13-52 (sites 4.5 mm apart; Fig. 6) or when a lateral reach (site 42) was paired with a jaw opening (site 52) in case G14-10 (sites 5.7 mm apart; Fig. 6). Both movements evoked from forelimb and face representations remained unchanged, and there was no difference between individual movements pretested and tested during dual stimulation. It must be emphasized that all these sites were quite distant from each other (4.5–7 mm). We do not have examples for dual stimulation of PPC sites evoking forelimb and hindlimb movements.

Our analysis of movements evoked from PPC domains (Figs. 7 and 8) supports the following conclusions. First, stimulation of two domains representing (two different) conflicting movements most often led to mutual interference (inhibition or suppression) with one movement winning out (prevailing). The trajectory of the resulting movement could be either shorter and of the average speed of both movements, or similar to the pretested prevailing movement. Less often, stimulation of PPC sites in different domains evoked a blend or an alternation of movements. Second, stimulation of two sites in the same domain resulted in the same kind of movement, with a longer trajectory and a higher speed. Third, costimulation of two domains with each representing the movement of a different body part did not produce any obvious interaction. Fourth, disruptions of movements evoked from two costimulated cortical sites in PPC were greater (12 of 13 pairs in PPC, 92.3%) than those evoked from M1 (14 of 41 pairs in M1, 34.1%; Table 2). Thus sites in PPC were more likely to interact in an antagonistic way than sites in M1.

Movements Evoked from Paired M1-PPC Sites

In two galagos, G14-10 and G15-03, we compared the movements evoked from sites in M1 and PPC when they were stimulated separately and together. Twenty-six pairs were studied, 12 in case G14-10 and 14 in case G15-03 (Table 1). The distributions of these sites on the surface outlines of M1 and PPC are shown in Fig. 9.

Fig. 9.

Distributions of stimulated sites on a surface view of primary motor area (M1) and posterior parietal cortex (PPC) in cases G14-10 and G15-03. Intracortical microstimulation (ICMS) responsive sites are marked by closed circles and unresponsive sites by open circles. Paired ICMS sites described in text are identified by large closed circles and site numbers. Zones of sites with different movement thresholds are shown in different shades of gray. Thick black lines indicate approximate borders of M1, and thin black lines indicate approximate borders of complex movement domains. Motor cortex is bordered by premotor cortex (PMD) and area 3a. FSa and FSp, anterior and posterior frontal sulci; IPS, intraparietal sulcus; PPCr, rostral PPC.

Stimulation of sites representing different movements.

Most of the studied pairs of sites, with one in M1 and one in PPC, evoked different forelimb movements when sites were individually microstimulated. The sites in each pair were 10 mm or more distant from each other. Two different outcomes occurred during costimulation of sites in M1 and PPC. When different arm movements related to the stimulated domains (e.g., hand to body and reach), the resulting movement was a blend of two participating movements (35%) or only the M1 movement occurred (65%), although less strongly (due to the conflict) or stronger (due to inhibition of the competing PPC site). Contrary to the previous cases with dual stimulation of pairs of sites in M1 or in PPC, costimulation of sites with one in M1 and another in PPC did not evoke an alternation of the two movements. Examples of combined movements obtained from pairs of sites 8+6 in case G15-03 and sites 38+36 and 45+42 in case G14-10 are illustrated in Fig. 10. In all these examples, the movement evoked by the dual stimulation was clearly a blend of individually evoked movements (Fig. 10, A–C). Thus the main characteristics of each single site movement were preserved and combined with another to produce a new movement. In case G15-03, the trajectory of the new movement was almost exactly halfway between trajectories of movements evoked from two individually stimulated sites (Fig. 10A). For both pairs of sites in case G14-10 (Fig. 10, B and C) the first phase of the new combined movement followed the trajectory of the movement evoked from the PPC site and the second phase followed the trajectory of movement evoked from the M1 site. For example, a grasp-with-hand-to-midline movement (site 38) paired with reach-up movement (site 36), resulted in a reach up (characteristic of the PPC site) followed by a grasp with hand to midline (characteristic of the M1 site). Similarly, pairing site 45, representing reach to midline, with site 42, representing reach forward, resulted in the movement that first followed reach-forward movement (characteristic of the PPC site) and then reach to midline (characteristic of the M1 site). The speeds of these new movements were the average of the speeds of both individual movements (Fig. 11, A and B).

Fig. 11.

Speed profiles of movements evoked by dual stimulation of sites in primary motor area (M1) and posterior parietal cortex (PPC) that represent different movements (A–E). Trajectories of movements during single and dual site stimulations are illustrated in Fig. 10. Speed was measured in 33-ms increments. The profile line for the costimulation is dashed. Arrowhead marks start of stimulation.

For other pairs the simultaneous stimulation of sites in M1 and PPC forelimb domains led to competition between movements, with one movement being suppressed (65%; Table 2). In case 14-10, a site in M1 (site 15) representing forelimb retraction was paired with a site in PPC (site 36) that represented reach upward. Simultaneous stimulation of these sites resulted in a forelimb retraction that was characteristic of the M1 site (site 15). The trajectory of this movement, although shorter, followed closely the trajectory of movement evoked from M1 when it was stimulated alone (Fig. 10D). This result, as well as similar results from another pair of sites in the same case (sites 44+42; Fig. 10E), indicates that one site (the M1 site) was able to suppress the function of the other (the PPC site). Because the speed of the expressed movement was lower than the speed of the single site-evoked movement (Fig. 11, C and D), the PPC site appeared to have a weaker suppressive effect on the M1 site. Simultaneous stimulation of two other pairs of M1-PPC sites, such as sites 4+36 (case G14-10) and 26+21 (case G15-03), demonstrated an even stronger suppressive effect for one of the competing sites. During dual stimulation of these sites, the resulting movements had longer trajectories (Fig. 10, F and G) and were faster (Fig. 11E) compared with their characteristics during pretesting (individual stimulation). Thus the suppression was so complete that the corresponding suppression from the other site was below the levels that occurred when only one site was stimulated. In looking at the competition between M1 and PPC sites, it appears that most often the movement evoked from M1 was the winner. In some instances, the prevailing of M1 movement might be due to the proportionately higher stimulation current applied to M1 domain compared with PPC. However, this cannot be the explanation for M1 domination in cases when both M1 and PPC domains were stimulated at thresholds (Fig. 10G, sites 26+21) or the “losing” PPC domain was stimulated at proportionately higher current than M1 (Fig. 10D, sites 15+36).

Stimulation of sites representing similar movement.

We studied four pairs of M1+PPC sites that evoked matching movements when electrically stimulated. In case G14-10, individual stimulation of either site 37 in M1 or site 36 in PPC (~12 mm apart) both evoked a reaching-up movement. Simultaneous stimulation of these two sites evoked a similar reaching movement, but with much longer trajectory and much faster speed than those of individual movements (Fig. 12, A and B). The maximum speed of the movement evoked by simultaneous stimulation of sites 37 and 36 was almost exactly the sum of speeds of the individual movements (site 37 speed, 21.2cm/s; site 36 speed, 24.4 cm/s, sites 37+36, 45.5 cm/s). Similar results of simultaneous stimulation of two sites in the functionally matched domain were observed for three other pairs of sites costimulated in case G15-03 (not illustrated). Thus stimulating PPC domain while stimulating a functionally matched M1 domain produces an additive effect, similar to a stronger electrical stimulation of the M1 domain.

As expected, all movements of the hand evoked by single and dual stimulation followed a typical bell-shaped speed profile, in which the peak speed was roughly linearly correlated with length of the hand movement (see also Bizzi and Mussa-Ivaldi 1998; Flash and Hogan 1985). Generally, movements evoked from PPC were about 33 ms delayed compared with movements evoked from M1. Our analysis of the movements evoked from the simultaneously stimulated forelimb domains in M1 and PPC shows that these movements were initiated at the same time as movement evoked from the M1 site when it was stimulated alone. Thus the new movement started usually within 66 ms after stimulus onset. When M1 and PPC sites evoking different movements were stimulated at once, the speed of the resulting movement averaged (sites 38+36 and 45+42; Fig. 11, A and B) or was lower (sites 15+36 and 44+42; Fig. 11, C and D) or higher than (sites 4+36; Fig. 11E) the speed of individually evoked movements. As for M1+M1 and PPC+PPC (see above), the speeds of same or similar movements evoked during paired stimulation of sites in M1 and PPC were higher compared with speeds of individually evoked movements (sites 37+36; Fig. 12B).

Stimulation of sites for forelimb and other body part movements.

In case G14-10, costimulation of a site in the PPC forelimb representation (evoking a lift with a grasp) and a site in the M1 face representation (evoking a grimace) at the same time did not reveal any visible interaction between those sites. As expected, their parallel stimulation resulted in exactly the same movements as movements evoked by stimulation of each site separately. Similarly, in case G15-03, costimulation of sites for eye blink (M1) and reach (PPC) evoked two independent movements of an amplitude similar to that of individually evoked movements. Thus, when M1 and PPC domains for different body parts are costimulated (e.g., arm and face), both movements of unchanged magnitude occur.

Convergence of Movements to an Endpoint

Although the starting position of the hand was controlled and kept similar over the course of our experiments, small spontaneous movements occurred over time. This resulted in small changes in starting hand position and in the trajectories of movements evoked from the same sites during the experiment. Thus the evoked movements were somewhat dependent on the starting point of the hand. However, regardless of its starting point for each stimulated site, the hand almost always moved toward the same final region in space. In this regard, we analyzed and compared trajectories of individual and paired movements evoked from sites in PPC (Fig. 13, A and A′), as well as from sites in M1 and PPC (Fig. 13, B and B′, C and C′) in galago G15-03. In all three cases, competition between two movements resulted in one movement prevailing. For each analyzed pair, the endpoint of the dual stimulation movement trajectory closely approximated that of one of the two sites, even though the starting position of the hand was different. Thus hand positions at the start of these movements differed, but their ending points overlapped substantially. The neural mechanisms for such convergence of movements to a specific endpoint can be found elsewhere (i.e., Bizzi et al. 1982; Bizzi and Mussa-Ivaldi 1998; Feldman et al. 1998).

Fig. 13.

Examples of movements evoked by dual stimulation that converge toward the endpoint of one of the two movements evoked by single-site stimulation. Trajectories of individual and paired movements evoked from sites in posterior parietal cortex (PPC; A and A′), as well as from sites in primary motor area (M1) and PPC (B and B′, C and C′), when stimulated separately (green and blue lines) and concurrently (dashed red line). Spacing of the dots reflects the frame time (33 ms). A–C: starting points of individual and paired movement trajectories are superimposed (see arrowheads). A′–C′: somewhat different actual starting points of the same movement trajectories are marked with arrowheads. Note overlay (outlined gray oval) of endpoints of winning movement trajectories when the cortical site was stimulated individually (during pretesting) or together with another site evoking concurrent movement.

Movements Evoked from Paired Sites of Left and Right Hemispheres

In one squirrel monkey (SM13-52), we also concurrently stimulated cortical sites in left and right hemispheres to see how domains in two hemispheres interact during evoked movements. Three pairs of sites in M1 and five pairs in PPC were stimulated (see Table 1). Movements evoked from PPC sites in left and right hemisphere seemed to interact more than movements evoked from M1 sites in both hemispheres.

The stimulation of single sites in M1 chosen for dual stimulation evoked mostly movements of the hand and digits. Thus the slight lateral hand lift with digits extension was evoked from the left hemisphere and grasp from the right hemisphere, or the grasp movements were evoked from both hemispheres. Similar movements were seen when these sites were coactivated, and their strengths were approximately the same as observed during the individual stimulations. Thus left and right M1 representing distal forelimb movements seemed to primarily operate independently.

No visible difference between forelimb movements evoked from left and right PPC either by simultaneous stimulation or independently was observed for three of five pairs of PPC. Movements were elicited with no apparent change in their trajectories, amplitudes, or speed compared with the pretesting condition (not shown). However, in two other instances, the movements evoked during dual stimulation differed somewhat from the pretesting movements. For example, when a hand to body movement (right hemisphere) was paired with defensive movement (left hemisphere), during the dual stimulation the hand-to-body movement appeared to be larger, whereas the other movement did not change compared with the pretesting condition. During the concurrent stimulation of two other sites in the left and right PPC representing lateral forelimb movements, the movements remained the same, but their amplitude increased.

DISCUSSION

Previous studies have shown that 500-ms trains of electrical stimulation applied to single sites in M1, PMC, and PPC evoke complex movements in primates and that sites for the same or similar movements are grouped, forming cortical domains (Baldwin et al. 2017; Gharbawie et al. 2011a, 2011b; Graziano et al. 2002a, 2002b; Stepniewska et al. 2005, 2009a). To understand better how these functional domains within and between frontal and posterior parietal areas interact with each other in producing complex motor behaviors, we simultaneously stimulated two cortical sites within these domains to evoke movements. We compared these movements with movements evoked by stimulation of each site on its own. Our results show that interactions between forelimb domains in M1, PPC, and M1+PPC are complex and that they vary, depending on the cortical area and the movement domains involved in dual stimulation (Fig. 14). Results can be additive when they are from pairs of sites that are functionally matched across areas, or antagonistic when from sites in functionally nonmatching domains. The individually evoked movements can also blend into a novel movement during costimulation. Costimulation of two different forelimb domains in M1 usually results in a blend of movements, whereas in PPC, dual stimulation mostly evokes suppressive interaction between domains with one movement winning the competition. Stimulating pairs of domains involving different body parts in either M1 or PPC produces both movements independently. Costimulation of sites within functionally different domains in M1 and PPC evoked a movement that was typically the same as or similar to the movement seen during the individual stimulation of M1. These results are discussed below in light of related literature.

Fig. 14.

Comparison of classes of movements evoked by simultaneous stimulation of sites in nonmatching domains in primary motor area (M1), posterior parietal cortex (PPC), and M1+PPC. Note that the most common effect related to simultaneous stimulation of pairs of M1 sites is a blend of movements, whereas the most common effect related to stimulation of pairs of sites in PPC or M1+PPC is the suppression of one movement by the other.

Stimulation of Two Domains Within Either M1 or PPC

To control behavior, the brain relies on neuronal networks to process information arising from external and internal sources. Knowledge about the intrinsic and extrinsic connections of cortical areas such as M1, PMC, and PPC is fundamental to understanding the neural processing occurring within and between these areas (see Caminiti et al. 2017 for review). Our earlier studies of parietal and frontal cortex connections indicated that movement domains in each region have numerous interconnections with other domains within the region (Gharbawie et al. 2011a, 2011b; Stepniewska et al. 2009b). These results, together with related anatomical and physiological data, suggest that during natural movements, widespread motor cortical foci within cortical areas or regions are functionally linked (Amassian et al. 1995; Devanne et al. 2002; Huntley and Jones 1991; Sanes et al. 1995; Schieber and Hibbard 1993). Such clustered, horizontal connections within motor areas may be used for coordinating the activity of different movement representations for the execution of complex movements and the suppression of conflicting movements.

To determine how different movement domains interact, we optically imaged the evoked cortical activity within the PPC region while stimulating PPC domains (Friedman et al. 2014; Stepniewska et al. 2011). Surprisingly, the optically imaged activation resulting from electrical stimulation of PPC domains did not spread widely or randomly. Instead, it produced above-threshold activation only within or next to the stimulated domain in PPC and within matching domains in M1-PMC (see also Adelsberger et al. 2014; Logothetis et al. 2010; Safavi et al. 2018 for related results). Because the widespread intrinsic PPC connections are in contrast to the limited zone of activation as measured with optical imaging, we suggest that the excitatory connections between domains terminate mainly on inhibitory neurons, reducing the activity in nonmatching domains in the same (PPC) and other (PMC or M1) regions (see Kaas and Stepniewska 2016). Thus cortico-cortical inputs would act, directly and indirectly, to excite corticospinal neurons in the same locations and simultaneously interneurons whose effect is to inhibit other corticospinal output neurons (Schneider et al. 2002). The connections between domains may require such stabilizing mechanisms; otherwise, activity at a stimulated point would spread beyond a domain and recruit other domains that are inappropriate for the intended movement. However, domains within M1 especially may also excite one another, depending on their functional roles, and the local inhibitory neurons themselves may be inhibited. Thus the functional linking of cortical sites may involve not only a focus on inhibitory neurons but also the greater excitation of output neurons, depending on the function of domain.

In the present study we evaluated the hypothesis of mutual inhibition between functionally mismatched domains by stimulating two domains at the same time. Our results are largely supportive of this hypothesis. During dual stimulation of two nonmatching forelimb domains, one domain often dominated the other, leading to a generation of movement that was similar to the movement evoked from one of the stimulated domains, either without the other movement component or with this component appearing much weaker than during pretesting. Thus both movements were present, but one of them was stronger than or totally dominated the other movement, suggesting the partial or total suppression exerted by one site on the concurrent site. This result also suggested an asymmetry of the cortically mediated mutual inhibition, with one site having a stronger inhibitory effect than the other.

Interestingly, observed disruptions of movements evoked from two costimulated cortical sites in PPC were greater than those evoked from M1. Thus sites in PPC are more likely to interact in an antagonistic way than sites in M1. What is the function of such suppressive interactions between domains? We suggest that suppressive effects of stimulating two mismatched domains in the three cortical regions (M1, PMC, and PPC) are involved in promoting one behavior over others. This occurs within or across cortical regions. However, domains in M1 are also more likely to excite each other to produce combined movements.

As proposed by ethologists, many behaviors are likely to be under inhibitory control as part of a hierarchical mechanism to allow high-priority behaviors to be preferentially selected (McCleery 1983; Tinbergen 1951). In our model, the behavior selection (or decision process) starts in the rostral PPC, whose domains are dominated by sensory inputs from the visually driven caudal half of PPC as well as higher order sensory and multisensory areas. PPC domains would select one behavior over another as sensory stimuli drive one domain sooner, or greater, than the others, which are then suppressed by inhibition generated by projections from the more active domain onto inhibitory neurons in other domains. The output of the most active PPC domain selectively activates matching domains in PMC and M1, which are also under other influences and thus may alter the PPC decision and have somewhat different interactions between domains. The existence of mutual inhibition between different functional regions as a probable mechanism for decision-making has also been suggested by others, and the role of PPC in decision-making is well documented (Gold and Shadlen 2007; Katz et al. 2016; Kiani and Shadlen 2009). A decision that is the selection of one behavioral outcome over others could result from interactions between domains that are overall suppressive or facilitative to others to produce a dominant output, as in the “race model” of decision-making of Logan et al. (1984).

In this study, we evaluated effects of simultaneously stimulating pairs of sites in the same cortical region, M1 or PPC, that evoked the same or very similar movements, different movements of the same body part (forelimb), or different movements of different body parts (e.g., forelimb and face). Cortico-cortical connections within regions appear to be very dense within a functional domain, quite dense with nearby domains for the same body part, and less dense or sparse between domains involving different body parts, for example, face and forelimb (Gharbawie et al. 2011a; Stepniewska et al. 2009b). They may also be patchy, especially in M1, and thus variable at similar distances from the injection site. Overall, these connection patterns seem consistent with many of our results from costimulating pairs of sites within M1 or PPC, but only if one assumes that within-region excitatory connections variably excite inhibitory or other excitatory neurons. We suggest that, near the stimulation site, more of the horizontal intrinsic axons contact excitatory neurons while longer horizontal connections contact inhibitory neurons. When we used optical imaging to reveal cortical activation patterns in the regions of a single stimulating electrode, the activation was high in the cortex immediately around the electrode and in small patches in the surrounding cortex (domain satellites), but not more distantly within the same region (Friedman et al. 2014; Kaas and Stepniewska 2016). These results suggest that the horizontal connections within anterior PPC or M1 of primates vary not only in decreasing density with distance but also in functions, with the longer connections activating more inhibitory neurons and depressing overall activity.

For some pairs of electrodes that were very close together in our study, the movements evoked by each electrode were very similar, and when electrodes were stimulating together, the evoked movement was faster and of greater magnitude. Because the electrodes were likely in the same domain, this could simply be the result of activating more neurons in the domain or the same neurons at higher rates. In part, the greater activation could be due to overlapping fields of current spread when the distance between sites was less than 2 mm (Logothetis et al. 2010; Tehovnik et al. 2006). To the extent that more neurons were activated, their outputs to other cortical areas and subcortical targets likely would be summed. Thus the faster, greater movements conform to expectation. When stimulation of two electrodes sites in the same region produced two different movements, such as reaches in different directions, the costimulation of such sites often produced a compromised movement, such as a reach between the two separate targets. Although we treat this result as having been produced by two different domains for different reaching movements, a reasonable alternative is to consider that a larger reach domain maps different reach movements such that the target of the reach depends on where in the map the greater stimulation occurs. With two stimulating electrodes, the center of greatest activation may move to between the electrode sites. Both the eye movement domains in PPC and the frontal eye field (FEF) of galagos (Fang et al. 2005; Stepniewska et al. 2018; Wu et al. 2000) and other primates (Schall 2015; I. Stepniewska, R. M. Friedman, A. W. Roe, and J. H. Kaas, unpublished observations) contain maps of eye movements extents and directions, so the center of activation within the map determines the movement (e.g., Mirpour et al. 2010).

Costimulation of unrelated domains (e.g., forelimb and face) within M1 or PPC produced both movements independently, and little interaction was observed between these sites. This is not surprising since there are only sparse or no connections between representations of different body parts (e.g., face, forelimb, and hindlimb) within M1 (Fang et al. 2002; Huntley and Jones 1991), in contrast to the much more dense connections within the same body representation (e.g., forelimb). However, some suppressive effect of one site by the other site was seen during stimulation of forelimb and hindlimb (but not face) representations, with one of the movements being stronger and the other weaker than during pretesting. An explanation as to why leg movement sites had more influence on forelimb movement sites than face sites is not clear but may reflect greater functional coupling of the extremities. The connections that mediate these weaker interactions may relate to callosal connections in M1 (Gould et al. 1986) or to a mixture of crossed and uncrossed corticospinal projections of M1 (Lacroix et al. 2004). In addition, the “running” domain in medial PPC has sites where both forelimb and hindlimb movements are evoked (Stepniewska et al. 2005), and this domain has more connections that involve forelimb domains (Wang et al. 2018).

Stimulation of Domains in M1 and PPC Together

We have shown before that functionally matched domains in PPC, PMC, and M1 are densely connected and function in a series, from PPC to PMC to M1, with some PPC projections directly to M1 (Gharbawie et al. 2011a, 2011b; Stepniewska et al. 2009b, 2011, 2014a). M1 provides the major output to the spinal cord and is critical for movement production (Passingham et al. 1983; Stepniewska et al. 2014a). This functional relationship was supported by our inactivation experiments with muscimol injections (Stepniewska et al. 2014a) and cooling (Cooke et al. 2015). To study further the interaction between frontal and posterior parietal domains, we stimulated one site in a PPC domain together with another site in a M1 domain. Stimulation of matching domains in these two areas produced summation effects that were similar to those produced by stimulation of two sites in a single M1 or PPC domain. The amplitude of resulting movement was roughly twice as large and the speed twice as fast. Because the functionally related domains in PPC and M1 are strongly linked by excitatory connections (Stepniewska et al. 2011), the electrically excited PPC projections to M1 and the electrically excited neurons in M1 might combine to provide a stronger output to the spinal cord. Of course, the direct and indirect subcortical projections of PPC to the spinal cord may add to this enhanced spinal cord activation (Galea and Darian-Smith 1994; Nudo and Masterton 1990; Rathelot et al. 2017).

During costimulation of M1 and PPC sites located in nonmatching domains, the resulting movement was typically the same as or similar to the movement seen during the sole stimulation of M1. Thus, in competition between these two sites, the electrically stimulated site in M1 appeared to be a winner. However, the movement resulting from such competition could be shorter or longer than the movement evoked by only M1 stimulation. This suggests that the PPC domain has a weak suppressive or excitatory effect on the mismatched M1 domain. The connections that mediate these modifications of the effects of stimulation are uncertain, but we know that PPC domains project to their matching M1 and PMC domains and not to mismatched domains (Stepniewska et al. 2009b). In addition to the activation of matching domains in M1 and PMC, PPC domain could also activate connections between domains in frontal cortex that added to, or subtracted from, the activity in M1 domain, thus altering its effectiveness. Pathways involving subcortical structures may also be responsible.

In some instances, the dual stimulation of nonmatching domains in M1 and PPC evoked movements with the main characteristics of each single site movement preserved, but combined to produce a new movement. The trajectory of this new movement was halfway between the trajectories of movements evoked from two individually stimulated sites, and the speed of these new movements was the average of speeds of both individual movements. Interestingly, the first phase of such new movement usually followed the trajectory of the movement evoked from the PPC site, and the second phase followed the trajectory of movement evoked from M1 site. The speed of the movement resulting from costimulation of nonmatching sites in M1 and PPC averaged the speeds of the movements evoked separately. In contrast, the speed of movements evoked during costimulation of sites in the same domain was higher compared with speeds of individually evoked movements.

Finally, our study on simultaneous stimulation of a few pairs of sites in the left and right hemispheres indicates that they largely fail to interact. However, two sites may modify each other somewhat, and movements evoked from PPC sites in the left and right hemisphere did seem to interact more than movements evoked from M1 sites of both hemispheres. More results are needed, but these limited results are consistent with stronger interhemispheric connections of forelimb representation of PPC compared with M1 (Gould et al. 1986; Killackey et al. 1983).

In summary, based on connections between different domains within a region, blending seems to be a prominent effect of domains interaction within M1, and mutual inhibition is a major outcome of domains interaction in PPC. Thus, in PPC (and less often in M1), domains controlling different movements are connected by intrinsic horizontal connections that likely focus on inhibitory neurons, decreasing overall activity. Such functional linking may promote a dominance of one movement domain in competition with another for initiating different, often incompatible behaviors. Therefore, most evoked behaviors reflect competition between parallel parietal-frontal networks. Within a parietal-frontal network, functionally related domains in posterior parietal and frontal cortex are densely connected and function in series from PPC to PMC and M1 (Stepniewska et al. 2011), with M1 providing a stronger output to the spinal cord, which is critical for complex movement production.

GRANTS

This work was supported by National Institutes of Health Grant R01 EY02686.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

I.S. and J.H.K. conceived and designed research; I.S., R.M.F., and D.J.M. performed experiments; I.S. and D.J.M. analyzed data; I.S., R.M.F., D.J.M., and J.H.K. interpreted results of experiments; I.S. prepared figures; I.S. drafted manuscript; I.S., R.M.F., D.J.M., and J.H.K. edited and revised manuscript; I.S., R.M.F., D.J.M., and J.H.K. approved final version of manuscript.