Inactivation of the Dorsal Premotor Area Disrupts Internally Generated, But Not Visually Guided, Sequential Movements

Abstract

As skill on a sequence of movements is acquired through practice, each movement in the sequence becomes seamlessly associated with another. To study the neural basis of acquired skills, we trained two monkeys (Cebus apella) to perform two sequential reaching tasks. In one task, sequential movements were instructed by visual cues, whereas in the other task, movements were generated from memory after extended practice. Then, we examined neural activity in the dorsal premotor area (PMd) and the effects of its local inactivation during performance of each task. Comparable numbers of neurons in the PMd were active during the two tasks. However, inactivation of the PMd had a marked effect only on the performance of sequential movements that were guided by memory. These results emphasize the importance of the PMd in the internal generation of sequential movements, perhaps through maintaining arbitrary motor-motor associations.

Significance statement: The dorsal premotor cortex (PMd) has long been thought to be a critical node in the cortical networks responsible for visually guided reaching. Here we show that PMd neurons are active during both visually guided and internally generated sequential movements. In addition, we found that local inactivation of the PMd has a marked effect only on the performance of sequential movements that were internally generated. These observations suggest that, although the PMd may participate in the generation of visually guided sequences, it is more important for the generation of internally guided sequences.

Figure 1.

Task and cortical maps for monkey S. a, Repeating task, targets, and movements for sequence 5-3-1. b, Repeating task, targets, and movements for sequence 1-2-4. c, Lateral view of cebus brain. Dashed lines indicate the M1–PMd border and the pre-PMd–PMd border. PS, Principal sulcus; ArS, arcuate sulcus; CS, central sulcus; IPS, intraparietal sulcus; LS, lateral sulcus; R, rostral; L, lateral. d, MR image after the chamber implantation for monkey S. The white dotted circle indicates the chamber outline. e, Intracortical stimulation map from monkey S. Letters indicate the movements evoked at each site: S, shoulder; E, elbow; W, wrist; D, digit; F, face; T, trunk. f, Intracortical stimulation thresholds are indicated by the filled circle size. g, Penetration sites for single-unit recordings. h, Muscimol injection sites in the PMd (sites 1–10) and M1 (sites 11 and 12). * indicates that site 2 was injected twice. Black numbers, Injection sites with a significant effect; gray number, injection site in which no significant effect was observed.

Figure 2.

Neural activity in the PMd during the two tasks. a, Activity of a Repeating enhanced neuron. The activity was enhanced for movements from target 1 to target 5. The increase was present after contact with target 1 (right; MI = 0.9) and before the contact with target 5 (left). This neuron displayed only modest or no changes in activity when the same movements were made during the Random task (t test, p < 0.001 for all moves). b, Activity of a Random enhanced neuron. The activity was enhanced for movements from target 3 to target 1 (right; MI = −0.6). This neuron displayed only modest or no changes in activity during the Repeating task (bottom; t test, p < 0.001 for move 1-5, p = 0.01 for move 5-3, p < 0.001 for move 3-1). The movement performed is indicated by the symbols and numbers at the top of each column. Each tick mark represents a spike. The rasters and histograms are aligned on target contact. The rasters show 10 trials. Histograms include all correct trials. Histogram bin width, 20 ms.

Figure 3.

Reaching end points before and after muscimol injection. Left, Preinjection; right, postinjection. E_A, Accuracy errors; E_D, Direction errors; gray dots, correct response; black dots, error response. The muscimol injection was placed at site 2 in monkey S (Fig. 1 h). a, End points for move 5 to 3. b, End points for move 3 to 1. Percentage of trials ending in each target are given below the targets. Touches between targets were counted as touches to the closest target. *p < 0.05.

Figure 4.

Effects of PMd inactivation. a–c, The performance data for an injection at site 2 in monkey S (Fig. 1h). a, Error rate in the Random task. PMd inactivation did not have any effect on number of errors (χ² test, p = 0.992 for move 1-5, p = 1.000 for move 5-3, p = 1.000 for move 3-1, df = 1; others, t test, p = 0.415). b, Error rate in the Repeating task. After the muscimol injection, the number of errors increased dramatically with all movements in the Repeating task (χ² test, p < 0.001 for all moves, df = 1). c, Predictive responses. The percentage of predictive responses decreased significantly after the muscimol injection (χ² test, p = 0.006 for move 1-5, p < 0.001 for move 5-3, p = 0.014 for move 3-1, df = 1). d–f, Population data for 14 effective injections at 13 cortical sites. d, Average error rate in the Random task. PMd inactivation did not have an effect on the number of error responses in the Random task (paired t test with 14 injection experiments, p = 0.260 for move 1-5, p = 1.000 for move 5-3, p = 0.273 for move 3-1, p = 0.753 for others, df = 13). e, Average error rate in the Repeating task. The error rate in the Repeating task increased significantly after the PMd inactivation (paired t test, p = 0.015 for move 1-5, p = 0.03 for move 5-3, p = 0.014 for move 3-1, df = 13). f, Average predictive responses. The percentage of predictive responses decreased after localized inactivation of the PMd (paired t test, p = 0.004 for move 1-5, p = 0.004 for move 5-3, p < 0.001 for move 3-1, df = 13). Other, Movements that were not a part of the Repeating sequence (5-3, 3-1, or 1-5). *p < 0.05.