Basic approach, database, and activity at rest

We studied the single-unit activity of neurons sampled from connected regions of the GPi and VLa. The spiking activity of isolated single units was recorded from both areas simultaneously in macaque monkeys while the animals performed a two-choice reaction time reaching task for food reward [48; 49]. In short, animals were trained to hold their left hand at a home-position located immediately to the left of the animal at waist height. Two visual targets were located 14 cm apart at shoulder height on a vertically mounted response panel positioned 30 cm in front of the animal. After a randomized home-position hold period, one of the targets (chosen pseudorandomly trial-to-trial) was lit and the animal had to move the left hand swiftly to that target (distance: approximately 37 cm) and then hold there to obtain a reward. The right arm was restrained in a padded splint throughout a data collection session. Multiple microelectrodes or 16-contact linear probes were positioned acutely in the arm-related regions of both nuclei [50].

We used a combination of electrophysiologic techniques to ensure that recordings were obtained from anatomically connected regions of GPi and VLa. Consistent with previous observations [34; 51], the location of the VLa nucleus was distinguished from surrounding thalamic regions—and from the cerebellar-recipient ventrolateral posterior nucleus (VLp) in particular—by the presence of a short-latency inhibition of spiking in response to electrical stimulation of the GPi and the absence of short-latency excitation in response to stimulation of the superior cerebellar peduncle (SCP, Fig 1B). This electrophysiologic localization of the VLa/VLp border was performed during initial mapping studies prior to formal data collection. Across multiple parallel electrode tracks, thalamic neurons inhibited by GPi stimulation (yellow tick marks in Fig 1A, examples 1 and 2 in Fig 1B) were always encountered at locations that were dorsal and anterior to neurons excited by SCP stimulation (green tick marks in Fig 1A, example 4 in Fig 1B). In most cases a narrow border (<0.5 mm) separated the deepest GPi-inhibited neuron and the dorsal-most SCP-excited cell (Fig 1A), and the location of this border aligned closely across adjacent electrode tracks. Coincident inhibition from GPi stimulation and excitation from SCP (e.g., Fig 1B3) was observed only at a small number of thalamic locations, and those were located exclusively in the narrow border between GPi-inhibited and SCP-excited regions (red tick mark in Fig 1A, example 3 in Fig 1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1.

(A-B) Exemplar results from microelectrode mapping in the vicinity of the VLa thalamus. Single units encountered along parallel electrode trajectories were classified as being located in striatum (light gray tick marks), Rt (dark gray), or the VL thalamus. Neurons in VL thalamus were further classified as VLa neurons if they were inhibited by stimulation (“stim.”) of the GPi and not excited by stimulation of SCP (yellow tick marks). Fig 1B1-2 shows overlaid peri-stimulus raw waveforms (left) and raster plots of sorted spikes (right) from two example locations in VLa (locations 1 and 2 in Fig 1A and 1B) at which GPi stimulation evoked a pause in neuronal activity and SCP stimulation had no effect. Neurons in VL were classified as VLp neurons if they were excited by stimulation of SCP and did not respond to stimulation of the GPi (green tick marks in Fig 1A; e.g., location 4 in Fig 1A and 1B). Neurons located at the border between VLa and VLp occasionally responded to stimulation of both GPi and SCP (red tick mark, location 3 in Fig 1A and 1B). Neurons that did not respond to stimulation were classified as VLa neurons (black tick marks Fig 1A) only if they were located antero-dorsal to the VLa/VLp boundary and postero-ventral to the Rt. (C) Regions of the GPi and VLa that belong to the arm-related BG-thalamic circuit were identified by testing for short-latency effects of electrical stimulation in arm-related areas of primary motor cortex. GPi neurons were included if sampled from regions at which stimulation of motor cortex (time zero) evoked a triphasic response at short latency (top, raster plot of sorted spikes). VLa neurons were included if sampled from regions at which stimulation of motor cortex evoked a pause or burst of activity at short latency (middle and bottom panels, respectively). BG, basal ganglia; CA, caudate; GPe, globus pallidus-externa; GPi, globus pallidus-internus; Rt, reticular nucleus of the thalamus; SCP, superior cerebellar peduncle; VA, ventral anterior nucleus; VLa, ventrolateral anterior nucleus; VLp, ventrolateral posterior nucleus; VMb, ventral medial-basal nucleus; VPM, ventral posterior-medial nucleus.

https://doi.org/10.1371/journal.pbio.3000829.g001

The subregions of GPi and VLa that compose the arm-related BG-thalamic circuit were identified by testing for short-latency effects of electrical stimulation in arm-related areas of primary motor cortex (Fig 1C). Previous studies have shown that stimulation of cortex can elicit complex triphasic responses in GPi neurons, via two- and three-synapse arcs through the BG [51; 52] (Fig 1C top), and that excitation of corticothalamic projections can elicit excitatory and inhibitory responses in VLa neurons [53] (Fig 1C bottom).

During subsequent data collection, thalamic single units were included in the database as VLa units if the electrode contact was located below the reticular nucleus of the thalamus (which is easily recognized neurophysiologically), above the VLa/VLp border (as defined above), and within 0.5 mm of a neuron affected by M1 stimulation. Single units sampled from the GPi were included if they were located in the dorsolateral portion of the nucleus and were encountered within 0.5 mm of a neuron affected by M1 stimulation. (See S1 Fig for anatomic locations of all included single units relative to observed electrophysiologic landmarks and inferred nuclear borders.)

For analyses of task-related changes in activity, a total of 209 single units met the criteria to be included as GPi neurons and 218 as VLa neurons (Table 1). These units were studied over the course of 126 ± 82 trials of the behavioral task (mean ± SD; mean 63 trials for each of two movement directions; minimum number of trials per direction: 16; minimum duration of recording: 174.7 s). As expected, the resting firing rate of GPi neurons was significantly higher than that of VLa neurons (Table 1; p = 1×10⁻⁸¹, rank sum test) [34], and those rates and the differences between neural populations were highly consistent for the two animals (Table 1; p = 1, rank sum test). Also, as expected, the action potentials of GPi neurons were short in duration as compared with those of VLa neurons (Table 1; p = 1 × 10⁻³³, rank sum test).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Database and basic properties of neurons.

https://doi.org/10.1371/journal.pbio.3000829.t001

The mean firing rate of most single units showed small but significant ramps during the start-position hold period (i.e., before presentation of the task’s go cue; p < 0.05, linear regression; 97% and 98% of GPi and VLa cells, respectively). This phenomenon is illustrated for exemplar GPi and VLa units in Fig 2A. The GPi unit’s firing rate increased slowly (4.3 spikes/s², p < 0.001) during the approximately 1.2-s hold period before appearance of the go cue (red tick marks in rasters, Fig 2A) whereas that of the VLa unit decreased slowly (−6.7 spikes/s², p < 0.001; Fig 2A). For both GPi and VLa unit populations, the observed slopes were distributed symmetrically around zero (means: −0.05 and −0.14 spikes/s² respectively; p = 0.29, rank sum test). Positively and negatively sloped ramps in activity were equally common (49% and 51%, respectively), and those fractions did not differ significantly between GPi and VLa neurons or directions of movement (p > 0.27, chi-squared test). These linear trends in delay period activity were taken into account by the algorithm used to detect peri-movement changes in firing rate, as described in the following sections.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2.

(A) Activity of one exemplar single unit sampled from GPi (left) and from VLa (right) aligned to the time of movement (“Mvt.”) onset (vertical black dashed line). Peri-movement spike-density functions (top) and rasters (bottom) show highly consistent changes in discharge rate around the time of movement onset (time zero). Vertical yellow and blue lines: the times of onset of increases and decreases in discharge detected relative to that unit’s baseline activity (sloping yellow trend line ± confidence interval). For raster plots, trials are sorted according to reaction time. Red tick marks: trial-by-trial times of go-cue presentation. (B) Spike-density functions of all single units studied sorted according to response form and response onset latency (earliest onsets at the top for each response form). Spike-density functions were z-scored relative to mean rate prior to go-cue presentation and displayed on a color scale. (C) Population-averaged spike-density functions for all GPi and VLa units (black) and for subpopulations with different response forms (as labeled). Shaded areas above/below the means reflect the SEM across units. Data and code to reproduce this figure can be found in https://doi.org/10.5061/dryad.0cfxpnvxm (Fig2_3_S8_S9.m). GPi, globus pallidus-internus; NR, no response; SEM, standard error of the mean; VLa, ventrolateral anterior nucleus.

https://doi.org/10.1371/journal.pbio.3000829.g002

Task performance

Both animals performed the behavioral task in a highly stereotyped fashion with short reaction times and movement durations (S2 Fig). Reaction times did not differ significantly between the two animals (F[1,452] = 0, p = 0.97, ANOVA) or between the two reach directions (F[1,452] = 0.2, p = 0.68). There was a slight (7 ms), yet significant, difference between the two animals in the effect of target direction on reaction times (animal × direction interaction; F[1,452] = 9.6, p = 0.002]. Movement durations were longer for reaches to the more-distant, right target than to the left (F[1,453] = 416, p = 4 × 10⁻⁶⁶). NHP G moved more slowly overall compared with NHP I (F[1,453] = 1,419, p = 1 × 10⁻¹⁴¹) and that slowing was more dramatic for the right target (animal × direction interaction; F[1,452] = 36.3, p = 1 × 10⁻⁹). Errors and outliers in task performance occurred at low rates in both animals (4.8% ± 1.2% and 3.4% ± 0.7% of trials in animals G and I, respectively; mean ± standard error of the mean [SEM]). Reach-related modulations in muscle activity (measured by electromyography [EMG], sampled during a subset of data collection sessions), were highly stereotyped across trials (S3 Fig). Importantly, reach-related changes in agonist muscle EMG began well before the overt onset of movement as detected mechanically and those onset latencies were similar for the two animals (−123 and −135 ms in NHP G and I, respectively; S3 Fig). The movement-related modulations in EMG differed quantitatively for the two directions of movement (widths of traces reflect SEMs in S3 Fig) but not in sign. The implications of these EMG results for the interpretation of single-unit activity are discussed in the following sections.

Movement-related increases in firing are common in both GPi and VLa

We examined the peri-movement activity of neurons in GPi and VLa to test for evidence of an inverse relationship in their rate changes, as predicted by the gating hypothesis. In these analyses, responses for each direction of movement were considered separately. Large proportions of both neural populations modulated their firing rates around the onset of reaches to the left or right target (Table 1). We constructed mean spike-density and ISI functions for each single unit separately for movements to left and right targets and then tested for significant changes in firing rate relative to that unit’s baseline activity (i.e., linear trend ± confidence interval [CI] of activity prior to go-cue presentation, yellow sloped lines in Fig 2A and S4 Fig). To avoid a bias inherent to spike-density functions (SDFs) toward detecting rate increases relative to rate decreases, we tested for peri-movement decreases in firing as increases in the mean ISI. (See S4 Fig for examples. See Materials and methods and S5 Fig for a simulation-based test for biases in the response detection algorithm used.) Fig 2A shows examples of the activity observed in individual GPi and VLa units. The GPi single unit showed a polyphasic change in firing that began with an increase in firing at −93 ms followed by a large decrease beginning at +77 ms relative to movement onset (yellow and blue vertical lines, respectively). The VLa single unit showed a monophasic increase in firing that first reached significance at −115 ms relative to movement onset. (The subsequent small but long-lasting depression in rate during movement did not reach significance.) Nearly all single units showed a significant peri-movement change in discharge for at least one direction of movement, with slightly fewer VLa units responding (99% of GPi neurons and 94% of VLa neurons; p = 0.02, chi-squared test).

The form and timing of neuronal responses differed widely between neurons (see Fig 2B and S4 Fig) and, occasionally, between directions of movement (see paragraph beginning “The observations outlined above…” and S6A and S6B Fig). First, we considered neuronal responses independently for each movement direction and tested for differences between GPi and VLa populations in the incidence of different forms of peri-movement activity (see Table 1 and Fig 2B). Monophasic changes in firing (composed of a simple increase or decrease in firing) were the most common change detected in both GPi and VLa, amounting to 66.7% and 88.9% of all responses detected, respectively. Monophasic responses were more common in VLa than in GPi (p = 2.8 × 10⁻¹², chi-squared test). Conversely, polyphasic responses—a series of one or more increase and decrease in firing—were more common in GPi than in VLa. Individual single-unit examples of each type of response are shown in S4 Fig, and all detected responses are shown sorted by response type and latency in Fig 2B.

Monophasic increases were the most common form of response both in GPi and in VLa (48.7% and 65.3% of responses detected, respectively; Table 1). When the individual phases of all detected responses were considered independently, increases in firing were also more common than decreases both in GPi and in VLa (61.5% and 68.8% of changes, respectively; p = 1.2 × 10⁻¹⁷, chi-squared test) with increases being nominally more common in VLa than in GPi (p = 0.03, chi-squared test). Simulations indicated that the observed high incidence of increases in both GPi and VLa was unlikely to be the byproduct of methodologic biases toward the detection of increases (see S5 Fig and Materials and methods).

Another way to compare the balance of task-related increases and decreases between GPi and VLa is to examine the mean firing rate across the two populations. Population averages combined across all responses types (all in Fig 2C) showed increases in firing rate during the peri-movement period for both cell types. This inclination toward increases was confirmed quantitatively by integrating changes in firing rate from baseline across the peri-movement epoch individually for each neuron. The mean of this integrated change was positive both for GPi and VLa (z = 7.37, p = 1.7 × 10⁻¹³, and z = 6.59, p = 4.7 × 10⁻¹¹ for GPi and VLa, respectively, rank sum test; Table 1).

The observations outlined above were confirmed in analyses that considered potential influences of the direction of movement on unit discharge. Although the majority of GPi neurons (109 out of 206) and some VLa neurons (56 out of 179) showed significantly stronger responses to one of the movement directions, movement direction did not affect the magnitude or timing of peri-movement activity across whole GPi and VLa populations (S6A Fig; z < 1.76, p > 0.07, and z < 1.0, p > 0.35 for GPi and VLa, respectively, rank sum tests performed separately for each time point, uncorrected for multiple comparisons). The form of an individual neuron’s peri-movement response (increase, decrease, or polyphasic) was typically consistent for reaches to the left and right targets (S6B Fig). This was true for both GPi neurons and VLa neurons (71% and 78% of neurons, respectively; shown by the color of bins along the matrix diagonals in S6B Fig; p < 10⁻¹⁶, chi-squared test for homogeneity computed separately for GPi and VLa distributions). Again, increases were more common than decreases and polyphasic changes were more common in GPi than in VLa populations (compare diagonals between matrices for GPi and VLa in S6B Fig; p = 2.9 × 10⁻⁶, chi-squared test). We found no evidence for directional biases in the incidence of different forms of responses (e.g., decreases were not more common for one direction of movement over the other). A comparison of integrated changes in firing rate further reinforced these conclusions. For both GPi and VLa populations, the polarity and magnitude of a neuron’s integrated change in firing for leftward movements was correlated with its change for rightward movements (S6 Fig. C; r > 0.6, p < 1 × 10⁻²⁵ for both GPi and VLa populations). Notably, few points fell in the upper left and lower right quadrants of S6 Fig. C reflecting the fact that few neurons were modulated reciprocally by the two movement directions or exclusively for one direction. Simple conceptions of the gating model predict that such relations with movement direction would be common in GPi. The polarities of integrated changes (i.e., increase versus decrease from baseline rate) and the degree to which polarities agreed between movement directions were distributed very similarly for GPi and VLa populations (S6D Fig; p = 0.68, chi-squared test). However, despite this general similarity of individual units' responses for left and right reaches, peri-movement activity did differ significantly between left and right reaches for substantial fractions of GPi and VLa neurons (53% and 31% of units, respectively; S6C Fig; p < 0.005, rank sum test).

In summary, the general distribution of different response types and the predominance of firing-rate increases were similar in GPi and VLa (see S6E Fig), contrary to the prediction from the gating hypothesis that this relationship would be reciprocal. There was general similarity in the form and the sign of responses detected in GPi and VLa with the most notable difference between populations being that polyphasic responses were more common in GPi than in VLa.

Response onset latencies in GPi and VLa are incompatible with gating and rebound

The gating theory predicts that decreases in GPi activity should precede and permit increases in VLa firing, whereas a rebound mechanism would also feature changes in GPi activity that precede those in VLa. To address these predictions, we compared the times of onset for all individual changes in firing detected in GPi and VLa single units. As is evident in both Fig 2B and Fig 3A, response onset times were distributed widely across the peri-movement period, and that was equally true for neural responses in GPi and VLa. Quantitative comparison, however, showed that, on average, GPi changes in discharge began later than those in VLa (median onset times: −47 and −90 ms for GPi and VLa populations, respectively; z = 3.31, p = 2.4 × 10⁻⁵, rank sum test; Fig 3A), in contrast to the relationship posited in the gating and rebound frameworks. The more specific gating prediction states that decreases in GPi activity should precede VLa increases and, similarly, GPi increases precede VLa decreases. In contradiction, the distribution of GPi decreases lagged in time behind that of VLa increases (median onset times: −27 and −88 ms for GPi and VLa populations, respectively; z = 3.19, p = 4.9 × 10⁻⁴, rank sum test; Fig 3B). Similarly, VLa decreases did not follow but rather preceded GPi increases (median onset times: −62 and −108 ms for GPi and VLa populations, respectively; z = 2.58, p = 0.01, rank sum test; Fig 3C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

(A) Cumulative distributions of onset latencies of all significant peri-movement changes in firing detected in GPi neurons (blue) and VLa neurons (purple). Responses in VLa precede responses in GPi by a median of 43 ms (**p = 2.4 × 10⁻⁵ rank sum test). (B) Response onset latencies of VLa increases (purple) lead GPi decreases (blue) by a median of 61 ms (**p = 4.9 × 10⁻⁴ rank sum test). (C) Response onset latencies of VLa decreases (purple) lead GPi increases (blue) by a median of 46 ms (**p = 0.01 rank sum test). Data and code to reproduce this figure can be found in https://doi.org/10.5061/dryad.0cfxpnvxm (Fig2_3_S8_S9.m). GPi, globus pallidus-internus; Mvt., movement; VLa, ventrolateral anterior nucleus.

https://doi.org/10.1371/journal.pbio.3000829.g003

Although the overall difference in distributions for GPi and VLa response latencies is inconsistent with the gating hypothesis, the latency distributions do overlap substantially (Fig 3), leaving the possibility that a subpopulation of early-onset GPi responses begin earlier than later-onset VLa responses and thus could contribute to their generation (or engage in “gating”). To estimate how common such cases might be, we compared the timing and sign of every individual response detected in the whole database of GPi units against every response detected in the VLa database (regardless of whether or not GPi and VLa units were recorded from simultaneously) and computed the fraction of comparisons that conflicted with the gating hypothesis with respect to the mentioned parameters. A large preponderance of these comparisons (96%, n = 178,846 of 187,154 comparisons) directly conflicted with the gating hypothesis because of the relative timing or signs of the responses (i.e., GPi onset precedes VLa onset by 0–50 ms and response signs are opposite). Thus, when the apparent large overlap in latency distributions illustrated in Fig 3 is additionally constrained by predictions regarding response signs, the possibility for a gating-like latency relationship was found in only 4% of all possible GPi-VLa neuron pairs. The potential presence of an early-onset GPi subpopulation is addressed in more detail in the next sections.

The latency distributions described here did not differ significantly between the two directions of movement for neurons in either GPi or VLa (z < 1.19, p > 0.08, rank sum test). In addition, the latency estimation algorithms used here showed only slight nonsignificant biases when applied to simulated neuronal responses that matched the metrics of real GPi and VLa neuronal responses (z < 0.19, p > 0.2, rank sum test; S7 Fig; see Materials and methods). Moreover, the small biases observed in the simulation results were toward longer lags for the detection of VLa responses relative to the detection of GPi responses—the converse of what was observed in the empirical latency results. Qualitatively similar results to those from the general comparison of latencies (Fig 3) were also found when the latency analysis was restricted to the preferred direction of single units with peri-movement activity that was significantly directional (S8A–S8C Fig).

As a further step to control for potential biases in latency estimation due, e.g., to differences between GPi and VLa populations in baseline firing rate and rate variability, we reestimated response onset latencies using an alternate approach based on the time at which an activity function crosses a fraction of the peak change (see Materials and methods). That approach resulted in slightly earlier-onset latencies overall (median latency shift: −14 ms), but the differences in latencies between GPi and VLa populations were fully consistent with those described above for the standard analysis (see S8D–S8F Fig). Thus, to summarize, the timing of changes in peri-movement discharge is not consistent with the idea that GPi activity triggers or gates task-related activity in VLa.

Finally, we looked for evidence that a subpopulation of early-onset GPi responses might lead and thereby gate subsequent VLa responses. For example, one might predict from the gating hypothesis that the magnitude of a GPi response is associated closely with the size of the response that it evokes milliseconds thereafter in VLa. Thus, GPi and VLa populations should show similar relationships between response magnitude and response latency if any such relationship existed. We therefore tested for relationships between response magnitude (measured here as mean change from baseline firing rate) and latency for all increases and decreases detected in GPi and VLa (S9 Fig). Contrary to predictions, however, the observed relationships differed markedly between GPi and VLa. In VLa, early latency responses were larger in magnitude than late onset responses as evidenced by the presence of significant negative correlations between latency and magnitude for both increase- and decrease-type VLa responses (Spearman p < 0.05) and significant 1-way ANOVAs across latency quartiles (F[3,248] = 4.3, p = 0.005 and F[3,112] = 4.3, p = 1 × 10⁻⁷, respectively). In GPi, however, no such relationship was found for increase-type responses (Spearman p = 0.95; F[3,308] = 1.5, p = 0.2), and nominally the opposite was found for decrease-type responses (Spearman p = 0.65; F[3,198] = 3.0, p = 0.03). In sum, the relation of response magnitudes and latencies does not support rebound- or gating-like mechanisms.

We next investigated the possibility that GPi-to-VLa communication might be evident in the activity of cell pairs sampled simultaneously from the two structures.

Correlated activity in GPi-VLa cell pairs is rare and unbiased

Both gating and rebound triggering of thalamic activity by the BG should produce strong correlations in the precise timing of spikes in GPi and VLa. Contrary to that expectation, we found little evidence for strong short-time-scale spike-to-spike interactions between GPi and VLa unit pairs (Fig 4A–4F). When we computed cross-correlation functions (CCFs) for pairs of spike trains sampled simultaneously from GPi and VLa, only 3.5% of the CCFs showed any statistically significant modulation (15 of 427 cell pairs; p < 0.05 relative to a 20-ms-jittered control; Table 2). Similarly, small fractions of CCFs reached significance when the analysis was restricted to a rest period during the start-position hold period (2.2% of pairs; 4 out of 183), or to the peri-movement period (5.2% of pairs; 9 of 172; Table 2). For the movement period, the fraction of significant correlations did not increase when correlations were computed separately for the two movement directions (4.1% of pairs; 14 out of 342), and there was no evidence that the occurrence of significant correlations was greater for one direction of movement over the other (6 out of 171 versus 8 out of 171; p = 0.59, chi-squared test). (Note that the number of pairs differs between time periods considered due to the strict selection criteria used to ensure adequate statistical power, as described in more detail in Materials and methods. The lower incidence of significant correlations in the direction-specific analysis may be related to the reduced number of trials.) Furthermore, the overall distribution of the peak absolute correlation values was not greater than the distribution of control peak absolute correlations taken from CCFs generated after jittering spike times within 20-ms time windows (Fig 4A–4C, uncorrected p-values: 0.975 for whole recordings, 0.675 for rest, and 0.998 for movement periods; one-sided permutation test). For none of the time periods considered did the rate of occurrence of significant correlations exceed by a significant degree the 5% rate of false positives detected in the control distributions. More broadly, population averages of the CCFs did not extend beyond the CIs from the control data (Fig 4D–4F, p > 0.05 relative to shuffled control).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4.

Absence of fast correlations between GPi and VLa spikes. (A-C) Histograms of peak absolute (“abs.”) CCF values between GPi and VLa units (black) for whole recordings (A), during rest (B) and during movement (C). Jitter control distributions (20-ms jitter intervals) are shown in gray. (D-F) Population average CCFs (black) and 95% confidence intervals (gray) based on the jitter control. (G-H) Low noise correlations between GPi and VLa during both rest (G) and movement (H). Noise correlations exceeding the 95% confidence limits are shown in orange (positive) and magenta (negative). Dashed vertical lines indicate the medians of the distributions. For both rest and movement, test distributions did not differ significantly from the control distribution (gray). No significant bias toward positive or negative correlations was found. Data and code to reproduce this figure can be found in https://doi.org/10.5061/dryad.0cfxpnvxm (Fig4_S10to13.m). CCF, cross-correlation function; GPi, globus pallidus-internus; VLa, ventrolateral anterior nucleus.

https://doi.org/10.1371/journal.pbio.3000829.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Cell-pair interactions between GPi and VLa.

https://doi.org/10.1371/journal.pbio.3000829.t002

Among the small number of individual CCFs that did exceed the threshold for significance, none showed features consistent with being the product of strong monosynaptic GPi-to-VLa inhibition. Contrary to the expectation that GPi-to-VLa CCF effects would be predominantly inhibitory (i.e., negative), roughly equal numbers of the CCF peaks detected as significant were positive and negative (8 versus 7, respectively; p = 0.79, chi-squared test; S10D–S10F Fig). Only two CCFs showed a long-lasting (>5 ms) reduction in VLa firing following the GPi spike (S10D Fig, first and third examples). Those two effects showed features consistent with a slow rhythmic GPi-VLa synchronization at 25 Hz (approximately 40-ms cycle length). Slow rhythmic synchronizations at frequencies <25 Hz were also present in some individual raw CCFs, but those effects were removed implicitly by subtracting the average jitter control CCF (see Materials and methods). The magnitude of individual CCF peaks correlated closely with the width of the CCF’s 95% CI (whole recordings: r = 0.82, p < 10⁻¹⁰; during rest: r = 0.62, p < 10⁻¹⁰; during movement: r = 0.61, p < 10⁻¹⁰), which reflects how noisy the estimate of the CCF was (S10G–S10I Fig). In other words, the narrower the 95% CI, the smaller the estimation noise and, incidentally, the smaller the detected CCF peak. This was true even for CCF peaks that were detected as significant (red X’s in S10G–S10I Fig), thus reinforcing the view that the CCF peaks detected here as significant may in fact have been false-positive noise events. Together, these observations suggest that short-time-scale spike-to-spike interactions between GPi and VLa unit pairs, if present, were too small and/or too uncommon to be detected with the methods used here.

Given these results, it was important to consider the sensitivity of our methods. To that end, we estimated how small of a known, experimenter-imposed, spike-to-spike correlation could be detected reliably in simulated spike-train pairs that matched the firing rates and recording durations of each empirical GPi-VLa pair. We also computed the upper bounds to each measured correlation (see Materials and methods for details). (These two measures were distributed very similarly within each analysis time period [whole recordings, rest, or movement; S11D–S11F Fig].) Not surprisingly, the minimum reliably detectable correlation (our measure of sensitivity) varied across the population of simulated pairs because of between-pairs differences in firing rates and duration of data collection. This led to a trade-off between the number of pairs considered and the within-pair sensitivity for detection of small correlations. Considering whole recordings, it was possible to detect correlations smaller than 20.1% of baseline rate in 90% of pairs (n = 388) and correlations smaller than 10.5% in 50% of pairs (n = 216) (S11A Fig). In the most sensitive 10% of pairs, it was possible to detect simulated correlations smaller than 6.1% of baseline. Even in this last subset, in which real correlations should have been most easily detected, the number and magnitude of empirically detected CCF peaks did not exceed what would be expected if the CCFs were composed of noise alone. Analyses applied to rest and movement epochs lead to similar conclusions (S11B and S11C Fig). Again, the magnitude of individual CCF peaks correlated closely with the minimum reliably detectable correlation (whole recordings: r = 0.98, p < 10⁻¹⁰; during rest: r = 0.97, p < 10⁻⁹; during movement: r = 0.98, p < 10⁻¹⁰), reflecting the noise in CCF estimation (S11D–S11F Fig), for all as well as for significant CCFs (red X’s).

In sum, short-latency cell-to-cell interactions were too small and/or too uncommon to be detected reliably using the approach applied here. These observations bring into question not only the standard gating and rebound hypotheses but also more fundamental assumptions about the mechanisms at play in GPi-VLa communication. We addressed this issue in more depth using a computational approach (see Simulations… section).

Next, we tested for slow trial-to-trial correlations in firing rate (“noise correlations”) between simultaneously recorded GPi-VLa cell pairs. This analysis was performed separately for rest and peri-movement periods. If gating was a common mechanism in GPi-VLa communication, then the majority of significant noise correlations would be expected to be negative whereas some versions of the rebound mechanism predict more frequent positive correlations due to the capacity of brief increases in GPi activity to effectively recruit prolonged rebound-supporting currents in VLa [4; 18; 19; 20]. We found that the overall distribution of noise correlations was not significantly different from the control distribution (Fig 4G and 4H; p = 0.07 during rest, p = 0.388 during movement; permutation test). Nevertheless, noise correlations did reach significance during rest for 5.4% of cell pairs (18 of 332 pairs; p < 0.05 relative to shuffled controls; Table 2) and during movement for 7.2% of cell pairs (27 of 372 pairs). Roughly equal fractions of those significant correlations were positive (orange lines, Fig 4G and 4H) and negative (magenta lines; 44% versus 56% during rest and 41% versus 59% during movement, respectively; Table 2). Also, between the recorded GPi neurons, cross-correlations and noise correlations were low (see S12 Fig). Examples of the largest three GPi-VLa noise correlations during movement are shown in S13 Fig. There was no evidence that the significant effects showed a bias in prevalence toward positive or negative correlations (p = 0.64 during rest, p = 0.34 during movement; chi-squared test).

Thus, neither of the tests for correlated activity in GPi-VLa cell pairs yielded evidence consistent with the gating or rebound hypotheses.

VLa activity decreases with bursts and pauses in GPi activity

Bursts and pauses are common features of GPi activity that could facilitate the transmission of information to VLa. The postsynaptic effects in VLa of GPi bursts and pauses could also be larger in magnitude and easier to detect than effects produced by single GPi spikes. We therefore estimated the influences of bursts and pauses in GPi unit activity on the firing rate of VLa neurons during the rest period (Fig 5). The gating hypothesis predicts that GPi bursts should be associated closely in time with reductions in VLa firing rate and GPi pauses with VLa increases. The rebound hypothesis predicts that GPi burst offsets should be followed by increases in VLa firing. We detected the occurrences of bursts and pauses in each GPi unit’s rest period activity separately using standard methods [54] and then averaged the firing rates of simultaneously recorded VLa neurons around the times of GPi burst onset, burst offset, and pause onset. The resulting burst-triggered averages (BTAs), burst offset–triggered averages (BOTAs), and pause-triggered averages (PTAs) of VLa activity were averaged across the populations of qualifying GPi-VLa pairs (Fig 5A–5C; Table 2; see Materials and methods for selection criteria and S14 Fig. for an example GPi BOTA VLa firing rate from one pair). Small but significant transient decreases in VLa population spike rate were evident following both the onsets and offsets of bursts. That decrease began at a longer lag following the onset of bursts (119 ms) than after their offsets (95 ms) with the difference (24 ms) equal to the observed mean duration of GPi bursts (24 ms). The presence of a decrease in VLa firing rate following the offset of GPi bursts is not consistent with the predictions of either gating or rebound hypotheses.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. VLa activity relative to GPi bursts and pauses.

(A-C) Population average VLa firing rates relative to GPi burst onsets (A), burst offsets (B), and pause onsets (C). The 95% confidence interval is shown in gray. In all three cases, VLa firing rates drop and stay low for several hundred milliseconds. (D) Population average cross-correlation of GPi pause onset times relative to the times of GPi burst offset (mean ± SEM). Pauses in GPi activity were more likely to occur immediately following the offset of a GPi burst. Data and code to reproduce this figure can be found in https://doi.org/10.5061/dryad.0cfxpnvxm (Fig5_S14.m). GPi, globus pallidus-internus; SEM, standard error of the mean; VLa, ventrolateral anterior nucleus.

https://doi.org/10.1371/journal.pbio.3000829.g005

Surprisingly, a deceleration in VLa activity was also associated with pauses in GPi activity (Fig 5C). The VLa decrease reached significance 50 ms after GPi pause onset. Given that GPi pauses had a mean duration of 195 ms, the decrease in VLa firing rate occurred during GPi pauses, not following them. Population-averaged cross-correlations of the times of pauses in GPi unit activity relative to burst offsets for the same unit (Fig 5D) showed a broad (>200 ms) period of negative correlation (i.e., reduced likelihood of a pause) at negative time lags, as would be expected because pauses in firing are unlikely to occur during bursts. The probability of a pause swung sharply to positive values at the time of burst offset and remained positive for >500 ms thereafter. Thus, bursts and pauses in GPi activity tended to occur together in that order, with decreases in VLa firing rates occurring following GPi bursts, during GPi pauses.

An analysis of individual GPi-VLa cell pairs provided results consistent with the population-level burst/pause analysis. Only small fractions of GPi-VLa pairs showed any significant change in VLa activity following GPi burst onsets, burst offsets, or pause onsets (Burst/pause influences; Table 2). Among those significant effects, decreases in VLa firing rate were far more common than increases, composing more than two-thirds of the significant effects for all events (Burst/pause influences; Table 2). Because of the small number of cases, however, those differences in prevalence were only nominally significant and only so for burst onset and offset (p = 0.04, 0.05, and 0.16 for burst onset, burst offset, and pause onset, respectively; chi-squared test).

Together, these results suggest that the activity of some GPi-VLa neuron pairs is coordinated such that burst-pause complexes in the GPi neuron’s activity are associated with long-lasting reductions in the firing rate of the VLa neuron. The characteristics of this phenomenon, however, are not consistent with predictions of either gating or rebound models.

Simulations suggest GPi-VLa communication scales with GPi synchrony and convergence

Given that GPi neurons project to VLa, it would be surprising if this pathway did not serve a functional role. Because we observed little impact of GPi activity on VLa in our empirical data, we turned to computational simulations to explore possible relationships between spiking in GPi and VLa (Fig 6). Specifically, we computationally generated sets of N GPi spike trains, with pairwise correlations c, for various choices of N and c. These N spike trains provided converging inhibitory inputs to a single model thalamocortical relay cell in which synaptic conductances were normalized by 1/N to maintain a constant average synaptic conductance across different choices of N (Fig 6A). (Homeostatic normalization of synaptic inputs similar to this is common in thalamocortical neurons [55].) Cross-correlations between spike times from randomly selected GPi input trains and the VLa neuron were computed for lags in the interval [−100 ms, 100 ms], and averages from 20 GPi neurons were collected over four separate runs for each (N,c) pair. Results showed a clear growth of the post-spike inhibition of VLa with increasing c (GPi pairwise correlation) when the number of converging GPi neurons (N) is held constant (Fig 6B). Importantly, when the within-GPi correlation is held constant, the post-spike inhibition of VLa decreased in magnitude as the number of converging GPi neurons (N) increased. Average correlations in the interval [0 ms,10 ms] were approximately proportional to [(N − 1)/N]c + 1/N as predicted on theoretical grounds (Fig 6C; see also Materials and methods). For large N and small c, the relative influence of individual GPi spike trains on VLa spiking became small owing to a large number of GPi spike trains contributing uncorrelated variability to the VLa membrane potential. Given a certain fixed amount of inhibition to each VLa neuron, these results highlight the importance of the degree of anatomical convergence from BG-output neurons onto recipient thalamic neurons, which is reported to differ markedly between homologous circuits in the songbird (in which N = 1; [25; 35]) and mammals (in which N may exceed 20). They also suggest that in contexts in which pairwise GPi cross-correlations are elevated (e.g., in parkinsonism [56; 57]), GPi outputs may exert a strong influence on subsequent VLa activity. Significant pairwise correlations in GPi spiking are rare, however, under baseline behavioral conditions in neurologically normal animals [56; 58; 59]. We confirmed that observation in the current data set by finding only 13 significant spike-to-spike correlations among 148 pairs of GPi neurons examined (S12A Fig). Out of these 13 significant correlations, three were classified as strong outliers (>10 × median absolute deviation away from the population median). Although the mean of all CCFs just exceeded the threshold for significance (S12D Fig, black line), that significance was lost after removal of those three outliers (S12D Fig, dashed line). No strong outliers were observed for the rest and movement periods (S12B and S12C Fig) or for any GPi-VLa correlation (Fig 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. GPi synchrony and anatomical convergence determine correlations between GPi and VLa.

(A) Computational model architecture. One VLa neuron is inhibited by N GPi neurons with pairwise correlation c. (B) Population average cross-correlations across (N,c) combinations. (C) Average correlations in the interval [0 ms, 10 ms] are similar to a linear approximation (dashed lines; see Materials and methods for details). Bars indicate the standard error of the mean of the 20 trials. Data and code to reproduce this figure can be found in https://doi.org/10.5061/dryad.0cfxpnvxm (Fig6.m). GPi, globus pallidus-internus; VLa, ventrolateral anterior nucleus.

https://doi.org/10.1371/journal.pbio.3000829.g006

In summary, the simulation results suggest that our empirical data were collected under conditions approximated by the model with parameters in the range of N ≥ 20 and c < 0.1 (i.e., high GPi-to-VLa convergence and low intra-GPi synchrony). Changes of those conditions—for example, increases in synchrony within GPi—may lead to efficient inhibition of thalamus by BG output.

Potential limitations and caveats

As discussed above, our results describe the dynamics of randomly sampled pairs of neurons in GPi and VLa, and they do not rule out the possibility that strong interactions, including gating or rebound, exist within tightly focused subcircuits connecting those nuclei. For example, it would be nearly impossible to detect by random sampling the very strong entrainment-like interactions observed in the hyperfocused pallido-thalamic circuit of the songbird [25; 35]. The anatomy of the mammalian GPi-VLa projection, however, suggests a far more branched organization containing a great deal of divergence and convergence [10; 11] in which it should be possible to study connected cell pairs by random sampling from GPi and VLa. At minimum, our results put a low upper limit (i.e., less than 15 in 427, see Table 2) on the probability of finding strong spike-to-spike cross-correlation effects, if any exist, in randomly selected GPi-VLa cell pairs. In addition, our results are inconsistent with the classic gating idea in which a coordinated drop in GPi activity is required to release thalamic activity and subsequent selection of action [40; 41].

In the comparison of onset latencies for GPi and VLa responses, a substantial degree of overlap in the distributions (Fig 3 and S8 Fig) leaves open the possibility for gating-like latency relationships in a minor subpopulation of GPi and VLa neurons. However, gating-consistent relationships of response sign and timing were found in only a small fraction (4%) of all possible combinations of GPi and VLa responses. In addition, inconsistent results for GPi and VLa were found when we tested for relations between response latencies and response magnitudes. Early latency responses in VLa were larger in magnitude than later responses whereas GPi responses did not vary in magnitude as a function of latency (S9 Fig). Furthermore, our study of cross-correlations and noise correlations in simultaneously recorded GPi-VLa pairs did not reveal any indication of gating or rebound mechanisms (Fig 4).

It is also important to acknowledge that we studied GPi-VLa communication during performance of a simple well-learned reaching task. It is possible that the influences of BG output on thalamic spiking could be stronger under more demanding or less-stereotyped behavioral contexts. For example, several lines of evidence suggest that BG-thalamic pathways drive behavioral variability or exploration during motor learning [89; 90; 91]. Other studies suggest BG involvement in the on-line modulation of movement vigor [44; 75] or the urgency to move [45]. A growing number of studies have concluded that the influence of BG output on a behavior becomes less important the more well-learned the behavior becomes [27; 77; 92]. It is possible that, under one or more of those less-stereotyped behavioral contexts, task-related activity in the GPi may adopt characteristics more capable of influencing VLa activity (e.g., larger firing rates, shorter latency responses to inputs, or more synchronization of spiking between GPi single units).

Conclusion

In conclusion, we found no evidence consistent with the idea that BG-output discharge gates thalamic discharge (“classic gating hypothesis,” [12]). The most likely alternative is that both pallidal and thalamic discharge may be driven by a third source [24; 25]. Layer 5 projections from motor cortex are a reasonable candidate for shared excitatory drive [63; 80; 81]. To the extent that they have been compared, all BG-thalamic projections in mammals appear to share similar anatomy and physiology [10]. Because of that, the present results have important implications for BG-thalamic communication in all functional circuits (e.g., in associative, oculomotor and limbic functional circuits [38]), not just the skeletomotor circuit. Our results are compatible with the idea that BG outputs may counterbalance cortical drive to thalamus. For example, increases in BG output may modulate or constrain the magnitude of thalamic changes in discharge, perhaps as a consequence of recent reward history [2]. Finally, subtle changes in BG-output synchrony may be a potent way to scale the effectiveness of inhibitory drive to thalamus.

Data collection

Offline analysis