Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Jun 3.
Published in final edited form as: Neuron. 2015 Jun 3;86(5):1174–1181. doi: 10.1016/j.neuron.2015.05.008

Antagonistic but not symmetric regulation of primary motor cortex by basal ganglia direct and indirect pathways

Ian A Oldenburg 1, Bernardo L Sabatini 1
PMCID: PMC4458709  NIHMSID: NIHMS689080  PMID: 26050037

Summary

Motor cortex, basal ganglia (BG), and thalamus are arranged in a recurrent loop whose activity guides motor actions. In the dominant model of the function of the BG and their role in Parkinson’s disease, direct (dSPN) and indirect (iSPN) striatal projection neurons are proposed to oppositely modulate cortical activity via BG outputs to thalamus. Here, we test this model by determining how striatal activity modulates primary motor cortex in awake head-restrained mice. We find that, within 200 ms, dSPN and iSPN activation exert robust and opposite effects on the majority of cortical neurons. However, these effects are heterogeneous, with certain cortical neurons biphasicly modulated by iSPN stimulation. Moreover, these striatal effects are diminished when the animal performs a motor action. Thus, the effects of dSPN and iSPN activity on cortex are at times antagonistic, consistent with classic models, whereas in other contexts these effects can be occluded or coactive.

Introduction

The basal ganglia (BG) are an interconnected group of subcortical nuclei that regulate movements and whose dysfunction contributes to multiple disorders (Albin et al., 1989; DeLong, 1990; Graybiel et al., 1994). Classical models of the motor BG describe a looped architecture in which motor cortex sends glutamatergic inputs to the striatum, the input stage of the BG, and is in turn influenced by the BG through inhibitory output onto thalamus. The two output pathways of the striatum, comprised of direct (dSPN) and indirect (iSPN) pathway striatal projection neurons, are thought to exert push-pull control over primary motor cortex (M1), by either increasing or reducing its activity to promote or suppress motor action. The anatomical substrates that mediate these antagonistic effects are thought to be the divergent GABAergic striatonigral and striatopallidal projections of dSPNs and iSPNs, respectively (Alexander and Crutcher, 1990; Deniau and Chevalier, 1985). The striatonigral projection inhibits the substantia nigra pars reticulata (SNr) whereas the striatopallidal projection inhibits the external segment of the globus pallidus (GPe). The GPe in turn inhibits SNr, making the net effect of iSPN activity to SNr excitatory (Gerfen et al., 1990). SNr provides GABAergic innervation of the ventrolateral thalamus (VL), which closes the loop via glutamatergic projections to cortex. This anatomical model explains the contributions of the BG to motor control, as well as the mechanisms by which symptoms of Parkinson’s disease are ameliorated by deep brain stimulation (Da Cunha et al., 2015) and is supported by lesion and pharmacological (Mink, 1996) as well as genetic and optogenetic (Bateup et al., 2010; Kravitz et al., 2010) studies.

Nevertheless, many features of this model have not been tested and are difficult to predict. The magnitude, kinetics, and homogeneity of a cortical response depend on many factors, including: the fraction of cortical activity that is driven by striatum-regulated thalamic inputs, the degree of tonic inhibition in the thalamus from ongoing SNr activity, and the speed with which cascading inhibitory networks disinihibit the thalamus and cortex. Many of these anatomical and functional parameters have not been determined, leaving fundamental aspects of the classic model of BG/cortical interactions untested and unconstrained.

Here we examine the control of cortex by striatum in awake, head-restrained mice. The effects of optogenetic manipulations of dSPN or iSPN firing on primary motor cortex were evaluated as mice performed a simple cued lever-pressing task for water rewards. At the level of populations of cortical neurons, our results generally support classic models of BG-cortical interactions. However, individual neurons can have heterogeneous, asymmetric, and context-dependent responses to manipulation of striatal activity, highlighting the existence of BG pathways by which dSPNs and iSPNs can have selective and non-antagonist effects on distinct cortical neurons.

Results

Studies of interactions between BG and cortex require analysis in awake animals as striatal activity is minimal under anesthesia (Mahon et al., 2006; Spampinato et al., 1986). Therefore, mice expressing Cre recombinase in either iSPNs (Adora-2A-Cre) or dSPNs (Drd1a-Cre) (Fig 1, S1A) and injected with Cre-dependent adeno-associated virus (AAV) encoding ChR2 were habituated to headrestraint. Mice were trained on a cued lever pressing task in which a motor action carried out shortly after an auditory cue led to a water reward (Fig 1A, S1BC, see methods). In trained mice, lever presses occurred preferentially after tones with press rate 2.75±0.53 fold higher in the reward period compared to similarly structured uncued periods (Fig 1A, p<0.01 Wilcoxon signed rank).

Figure 1. Channelrhodopsin mediated modulation of striatum.

Figure 1

Open in a new tab

A, Schematic of task design (top). A trial starts with an uncued 1.5–3 s withhold period (red). If the animal does not press the lever during this time, a 10 kHz tone is presented (vertical black line) which is followed by 1.5 s potential reward period (green). If the animal presses and releases the lever during this period, it receives a water reward (blue line). This is followed by inter-trial delay (3–8 s) during which presses are neither rewarded nor punished. bottom, Lever press rates during recording sessions (n=20, 8 mice) for periods of 1.5 s without lever presses (t= −1.5 to 0 s) that ended (t=0) with (black) or without (orange) the cue. Inset, finer timescale analysis (10 ms bins) shows that press rates diverge across conditions after ~50 ms.

B, Sagittal slices showing ChR2 expression (red) following injection of Cre-dependent ChR2-mCherry encoding AAV in mice that express Cre in iSPNs or dSPNs .

C, top/middle, Example raster plots and histograms of activity of highly-modulated units from iSPNChR2 (left) and dSPN-ChR2 (right) animals. Blue=473nm illumination. bottom, Histogram of IChR2 for recorded units. Red indicates statistically significantly modulated units (t-test, p<0.05, iSPN 35 of 76 units; dSPN 57/98).

D, Latency to modulation of striatal units. IChR2>0.75: iSPN n=7 units; dSPN n=8; IChR2 0.1–0.5: iSPN 106±44 ms, n=8; dSPN 125±16 ms, n=49; IChR2<−0.1: iSPN 144±48 ms, n=9; dSPN 250±58 ms, n=3. All units with latency < 500 ms are included.

E, ChR2-induced changes in behavior for iSPN-ChR2 (n=7), dSPN-ChR2 (n=8), or ChR2-negative control (n=3) mice. Relative lever press rates (left) and durations (right) are the ratios of each metric with and without stimulation (* indicates p<0.05, Wilcoxon signed rank).

Mice that reached behavioral proficiency were implanted with a fiber optic and analysis of the effects of ChR2 stimulation were examined on a recording rig. The stimulating laser was on or off continuously for each trial and switched to the opposite state such that transitions occured in intervals well separated (3–8 s) from the reward and at least 1.5 s before a tone. Multielectrode array recordings in striatum confirmed effective optogenetic manipulation (Fig 1C). The degree of modulation of each unit was calculated as:

IChR2=fon−fofffon+foff

with fon and foff corresponding to average firing rates with the laser on and off, respectively, during a 1.5s period prior to the delivery of the cue where the animal does not press the lever.

ChR2-stimulation modulated striatal neurons with IChR2 distributed over most of its −1 to 1 range. Optogenetic stimulation increased firing rates in 39% (30/76) and 87% (85/98) of units when activating iSPNs and dSPNs (Fig 1C), respectively, presumably through a combination of direct activation and network effects. In each condition, ~10% had IChR2>0.75 (iSPN experiments: 7 units; dSPN: 9). These putative ChR2-expressing units had low basal firing rates and responded with short latency to light. Units with intermediate activation had higher basal firing rates and responded more slowly (Fig 1D, S1DE). Significant inhibition of SPNs was rare following activation of dSPNs (4 units) and more common following iSPN activation (27 units) (Fig 1C). Such inhibition could result from SPN to SPN GABAergic synapses as well as from long-range circuit effects (see below).

SPN activity was modulated by the task. SPNs had high press related modulation indices (Ipress), calculated by comparing activity in ±0.25s around a spontaneous lever press to non-press periods (iSPN experiments: Ipress=0.21±0.04; dSPN: 0.26±0.07). Furthermore, stimulation of iSPNs and dSPNs bidirectionally modulated lever press frequency (ratio of frequency with light on vs. off: iSPN 0.45±0.09, n=7 mice, p<0.05; dSPN 3.1±0.66, n=8, p<0.05; Wilcoxon signed rank) whereas control mice showed no significant modulation (1.1±0.06 n=3). The duration of lever presses increased with activation of iSPNs but not dSPNs (iSPN: 6.3±2.9 fold change, p<0.05; dSPN: 1.2±0.27, n.s.; control: 0.93±0.07, n.s.; Wilcoxon signed rank; Fig 1E).

Effects of dSPN and iSPN activation on motor cortex

To determine the effects of striatal activity on cortex we inserted multielectrode arrays in the forepaw region of primary motor cortex (M1) contralateral to the lever and ipsilateral to the stimulated striatum (Fig S2A). The stereotaxic location of forepaw was confirmed via microstimulation in anesthetized mice (Fig S2B). Furthermore, activity in this area is necessary for the task as focal injection of GABA transiently impaired performance (Fig S2C) and is sufficient, using receiver-operator characteristic analyses, to predict the timing of spontaneous lever presses (area under curve=0.86±0.02, n=8 mice).

Firing rates of M1 neurons were compared with and without optogenetic stimulation during a 1.5s “baseline” period that ended with the tone and lacked lever presses, auditory cues or rewards. Consistent with classical models, activation of iSPNs reduced the firing rates of ~70% of units (Fig 2A, S2D): of 193 units (n=4 mice), the firing rates of 136 were significantly changed with 132 inhibited and 4 excited (p<0.05, 2-tailed t-tests on alternating trials). The population firing rate was reduced with a modulation index (IChR2) of −0.31 corresponding to a ~50% decrease (Fig 2B; p<0.0001; Matched pairs signed rank).

Figure 2. Antagonistic modulation of primary motor cortex by direct and indirect pathways.

Figure 2

Open in a new tab

A, Activation of iSPNs decreases (left) and dSPNs increases (right) firing rates in motor cortex. Example raster plots (top) and histograms (bottom) of activity of cortical units prior to and during optogenetic stimulation of striatum (blue).

B, IChR2 of cortical unit modulation with iSPN or dSPN stimulation. Red indicates statistically significantly modulated units (iSPN 136/193, 4 mice; dSPN 103/136, 4 mice; t-test, p<0.05).

C, Mean firing rate of cortical neurons at the start and end of ChR2-stimulation (blue) of iSPNs (left) and dSPNs (right). Gray: ± SEM.

D, Pseudocolored plots of firing of all units normalized to rates in baseline period and ordered by IChR2 (low to high). Blues/purples and yellow/red represent relatively decreased and increased rates.

Conversely, with optogenetic manipulation of dSPNs, activity increased in ~75% of M1 units (Fig 2A, S2D). Of 136 units (n=4 mice), 103 significantly changed firing rates with 100 excited and 3 inhibited (p<0.05 2-tailed t-test on alternating trials). dSPN activation significantly increased the population firing rate with a modulation index (IChR2) of 0.28 (Fig 2B; p<0.0001) corresponding to a ~80% increase. The average baseline firing rates in M1 were the same for iSPN and dSPN experiments (iSPN 9.6±0.99 Hz, dSPN 9.0±1.3, n.s. Mann Whitney). No robust, consistent change in the pairwise correlations across M1 units was observed due to activation of either pathway (Fig S2E).

Manipulations of iSPNs and dSPNs significantly modulated the majority of M1 neurons recorded. However, in each case a fraction of neurons were not significantly affected (iSPN experiments: ~30%; dSPN 25%) – percentages larger than expected false negative rates based on power analyses and confidence intervals (Fig S2D; supplemental methods), suggesting the existence of intermingled cortical cells whose activity is insensitive to the manipulations delivered to striatum.

Kinetics of striatal modulation of cortex

Modulation of cortex by striatum involves inhibition and disinhibition in a polysynaptic circuit that consists of cascading spontaneously active GABAergic projection neurons. Increasing the activity of downstream structures occurs via relief of tonic inhibition, a process whose kinetics is limited by the firing rates of intermediary neurons. We found that the latency for significant alterations in activity of M1 units by striatal activation was 123±7 and 169±21 ms following activation of dSPNs and iSPNs, respectively.

Unexpectedly, immediately following ChR2-activation of iSPNs average M1 activity increased before decreasing (Fig 2CD), and effect due to a transient increase in firing rates in a subset (59/193) of cells. These units responded at an intermediate latency (140±11 ms, Fig 3A). To identify the transiently up-regulated units, we calculated modulation index Iearly comparing the firing rates 0.5 s before and after laser activation and examined units with Iearly>0.1 or Iearly <−0.1 (Fig 3B). Units with Iearly>0.1, found in nearly all recordings, were transiently activated at both light-on and -off (Fig 3B, S3A–C). Within 0.5–1s these neurons decreased firing rates, such that the overall IChR2 was negative.

Figure 3. Transient activation of motor cortex by the indirect pathway.

Figure 3

Open in a new tab

A, Latency of cortical response to striatal activation. All units with latency less than 500 ms are included. dSPN: 123±7 ms, n=125 units; iSPN transient activation: 141±11 (n=90); iSPNs without transient activation: 169±22 (n=44).

B, Average firing of cortical units separated into those transiently inhibited (black) or excited (green) by iSPN activation (blue bar).

C Average firing of cortical units (left) reveals greater transient activity in superficial (green) than in deeper (red) cortical units. Iearly plotted as a function of depth from the pia (right).

D, Similar analysis as in C for manipulation of dSPNs.

Units with positive Iearly were detected at electrode sites shallower than those with negative Iearly (579±29 µm vs. 874±40, p<0.0001, Mann Whitney; Spearman’s correlation rs= −0.41, p<0.0001. Fig 3C, S3D). Conversely, average Iearly was positive (0.19±0.05, n=93) for shallow units (100–750 µm) and negative (0.19±0.05, n=83) for deep units (>750 µm), indicating that transient activation following iSPN stimulation is more likely in superficial cortical layers. Whereas a difference in Iearly was apparent as a function of depth following iSPN activation, no similar phenomenon was seen with dSPN activation (Fig 3D, S3E–H).

Effects of dSPN and iSPN activation on motor cortex during movements

We separately examined the effects of striatal manipulations on M1 during different aspects of the task, following the tone alone (i.e. when the animal failed to press the lever) and during spontaneous presses (uncued lever presses outside of the reward period). As in the baseline periods, ChR2-activation of iSPNs decreased the firing rates of M1 units at the time of an uncued press or in tone-only trials (Fig 4A, p<0.0001 Wilcoxon matched-pairs signed rank). However, the degree of inhibition was weaker in the ±0.25 s surrounding uncued presses than during the baseline (IChR2= −0.31±0.02 vs. IChR2 press= −0.06±0.02, p<0.0001, KW statistic: 512.7, Kruskal-Wallis with Dunn’s Multiple Comparison; 132/193 units inhibited in baseline vs. 47 during movement), but unchanged during toneonly trials (IChR2 tone-only= −0.24±0.02, p>0.05; 132 units inhibited during baseline vs. 118 during cue, Fig S4D). Trials containing both tones and presses (i.e. success trials) revealed an intermediate response to optogenetic stimulation (Fig S4AB).

Figure 4. Differential effects of striatal activation on cortex.

Figure 4

Open in a new tab

A–B, Average rates of cortical units (top) normalized to basal firing aligned either to the time of a tone in failure trials or of a spontaneous lever press outside of the reward period. Units with >30 spikes in the baseline periods for each event class (press or tone) were included. Trials with optogenetic with activation of iSPNs (A) or dSPNs (B) are in blue and without in gray. Shading: ±SEM. bottom, Individual units’ normalized firing rates presented as a pseudocolored plot (as in Fig 2D) and ordered by the press or tone modulation index (low to high), without (left) or with (right) optogenetic activation.

C, Average firing rates of units during the 1.5 s baseline, ±0.25 s around a press, or 0.5 s after a tone with (y-axis) and without (x-axis) iSPN (left) or dSPN (right) activation. Error bars=SEM. iSPN activation (left) decreased firing rates for Baseline (9.6±1.0 Hz off, 5.5±0.6 on, n=193, p<0.0001), Tone (11.5±1.2 off, 6.8±0.7 on, n=193, p<0.0001), and Press (16.1±1.4 off, 13.6±1.2 on, n=179 p<0.0001). dSPN activation (right) increased firing rates for Baseline (9.0±1.3 off, 12.7 ± 1.6 on, n=136, p<0.0001), but not Tone (14.8±1.8 off, 14.7±1.8 on, n=136, p>0.05), or Press (20.3±2.2 Hz off, 19.7±2.1 on, n=136, p>0.05). Wilcoxon matched pairs signed rank.

D, Ipress calculated with (IpressON) and without (IpressOFF) ChR2 stimulation of iSPNs (green) or dSPNs (purple) are strongly correlated (p<0.0001; Spearman’s rs: iSPN 0.49; dSPN 0.83).

E, Changes in the ability of an observer to identify movements based on total cortical activity were measured with ROC analysis. Resulting area-under-curve values with and without iSPN (left) and dSPN (right) activation for each recording session.

F, IpressOFF and IChR2 of cortical units are not correlated for iSPN activation (left, p>0.05; Spearman’s rs= - 0.003) but are highly correlated for dSPN activation (right, p<0.0001; Spearman’s rs= 0.710).

Unlike the indirect pathway and in striking contrast to the baseline period, ChR2-activation of dSPNs did not affect firing rates during uncued presses and tone-only trials (Fig 4B, S4E, IChR2 press=0.02±0.02, IChR2 tone-only=0.00±0.02, n=136 units, p>0.05, Wilcoxon matched-pairs signed rank). The lack of elevated firing rates was not due to a ceiling effect as the rates during tone-only trials, with or without light, were significantly less than during uncued presses and success trials (Fig 4C; tone only 14.8±1.8 Hz vs. press 22.6±2.4, p<0.0001 Wilcoxon matched-pairs signed rank).

In order to determine the effect of optogenetic manipulation on the dynamic activity of cortical units, indices corresponding to activity during aspects of the task were analyzed. Ipress was calculated as above and, as expected in primary motor cortex, individual units were strongly modulated during spontaneous presses (iSPN Ipress=0.31±0.02; dSPN 0.45±0.03; Fig S4F). Similarly, Itone and Isuccess were calculated for the activity in tone-only (i.e. failure) and success trials, comparing the baseline activity to that in 0.5s period after the tone (iSPN Itone=0.04±0.02, Isuccess=0.22±0.02; dSPN Itone=0.32±0.02, Isuccess=0.37±0.04; each >0 with p<0.05, Mann Whitney).

In the simplest analysis, the motor character, or tuning, of individual units is unaffected by manipulation of each pathway in the striatum: in both sets of experiments Ipress measured without and with striatal activation are correlated (Spearman’s rs: iSPN 0.49, dSPN 0.83, p<0.0001; Fig 4D). Thus, units that significantly changed activity at the times of uncued presses without optogenetic stimulation continued to do so with stimulation. Furthermore, Ipress of individual units generally increased with activation of iSPN and decreased with activation of dSPN, an effect that was also clear at the population level (Fig 4D, S4G).

Such changes suggest that the ability of an observer to predict the onset of a spontaneous movement based on activity in M1 is enhanced by activation of iSPNs and degraded by that of dSPNs. Indeed a population spike count threshold model revealed such effects when analyzed by receiver-operator characteristics (ROC). In this model presses are generated at periods of high population firing above a threshold (Fig S4H, see methods) with no time dependence. This model generated good predictions of movement onset with area under the curve (AUC) values of 0.83±0.02 and 0.90±0.03 for iSPN and dSPN experiments. Upon optogenetic activation of iSPNs, AUC increased in nearly every recording (to 0.92±0.01 with iSPN activation, 11 recordings, p<0.01 Wilcoxon matched pairs signed rank). Conversely, upon dSPN activation AUC decreased in every recording (to 0.81±0.04 with dSPN activation, 9 recordings, p<0.001; Fig 4E, S4I).

Given the observed changes in M1, we examined the possibility that BG exert selective control over distinct cells in motor cortex. For each unit we compared the modulation of firing during presses (Ipress) to its modulation by BG activation (IChR2). The degree of modulation of each unit by activation of the iSPNs was not predictive of the degree of modulation of the unit by spontaneous movements – i.e. Ipress and IChR2 showed no correlation (Fig 4F, Spearman’s rs= −0.003 n.s.). In contrast, activation of dSPNs increased the basal activity of neurons in M1 that were more active at the time of the press – Ipress and IChR2 were highly correlated (rs=0.71, p<0.01). In effect, dSPN activation preferentially modulates M1 neurons that are active during movements, a specificity that is not seen following activation of iSPNs, suggesting that the motor cortex neurons most sensitive to the activity of iSPNs are not the same as those most highly regulated by dSPN activity.

Discussion

Our results demonstrate that in habituated, head-restrained mice activation of the iSPNs and dSPNs suppresses and enhances, respectively, firing rates of units in motor cortex, consistent with classic models of BG/cortical interactions. However, the effects are spatiotemporally heterogeneous and three surprising findings emerge that are not immediately predicted by classic models. First, iSPN activation unexpectedly excites a subpopulation of superficial M1 cells such that both dSPN and iSPN activity can be at least transiently excitatory. Second, the task related activity of neurons that are highly sensitive to dSPN stimulation is different that that of neurons highly sensitive to iSPN stimulation. These two findings indicate that the subsets of neurons in primary motor cortex regulated by each pathway are at least partially non-overlapping, highlighting the existence of separate routes by which dSPNs and iSPNs can modulate cortical activity. Third, trained movements and cues reduce or prevent the effects of dSPN activation on motor cortex activity, but have relatively little influence over the effects of iSPN activation. These differences underlie the non-intuitive result that the ability of an ideal observer to predict the timing of spontaneous movements based on analysis of total activity in primary motor cortex is enhanced by iSPN and degraded by dSPN stimulation.

Classic Models

dSPN and iSPN activation caused opposite ~3 fold changes in spontaneous lever press frequencies and iSPN activation increased lever press duration, a freezing-like behavior. Such effects are consistent with classic models of direct/indirect pathway functions (Albin et al., 1989) and of Parkinson’s disease (Marsden, 1982) as well as with recent studies in mice (Bateup et al., 2010; Kravitz et al., 2010). Furthermore, activation of dSPNs and iSPNs respectively increases and decreases firing rates in primary motor cortex during periods when the mice are not exposed to any of the overt task features – i.e. no lever presses, cues, or rewards. The effects on M1 activity during this period are strong (~2 fold modulation), widespread (>70% units showing significant modulation), and consistent with predicted antagonistic effects of each striatal pathway.

Within the context of classical models, our results provide evidence in favor of assumptions about activity in the cortex-basal ganglia-thalamus recurrent loop that are often not directly stated but that nevertheless assumed. For example, in this model and in order for the dSPNs and iSPNs to bidirectionally modulate cortical activity, it is necessary that SNr output provide tonic inhibition of the thalamus that is significant but not saturated. Although SNr output neurons are tonically active, synaptic depression during maintained high frequency firing might diminish the inhibitory influence of BG output on thalamus. Furthermore, in order to translate changes in BG output into alterations of basal firing rates in cortex, thalamocortical projection neurons need to both supply sufficient ongoing activity to account for a significant fraction of cortical excitatory drive and be under the control of BG. Thus, when coupled with the behavioral effects described above, bidirectional modulation of basal firing rates in primary motor cortex by dSPN and iSPN activation supports the classic model of BG function and its foundational assumptions.

Beyond classic models

The simple classification of dSPNs and iSPNs as pro- and anti-kinetic pathways, respectively, does not fully account for the activities of these cells in behaving mice since neurons of both classes are active during both the initiation and suppression of movements (Cui et al., 2013; Isomura et al., 2013). Furthermore, in monkeys, BG activity is concurrent or delayed relative to movement initiation, suggesting a function in shaping but not necessarily initiating motor action and associated circuit activity (Aldridge et al., 1980; Hikosaka et al., 1989; Mink and Thach, 1991). Resolving these issues requires knowledge of the kinetics of effects of striatal activity on other brain structures. We find that activation of dSPNs or iSPNs modulates M1 activity with ~150ms latencies (average ~120 dSPN, ~165 iSPN), with some cells responding in less than 50 ms. This is slower than striatal modulation of SNr (Freeze et al., 2013), consistent with the presence of additional two synapses between SNr and cortex. Given the short latency of cortically-evoked action potentials in striatum (Koralek et al., 2012), a complete closed loop interaction from cortex to BG and back likely can occur in less than 200 ms. Such recurrent effects may explain the large fraction of striatal neurons that are inhibited by activation of iSPNs (Fig 1C), which likely indirectly suppresses corticostriatal projections and decreases striatal activity with a delay.

The results presented here also reveal a complex dynamic response in cortex to striatal activation that violates the predicted symmetric effects of dSPN and iSPN activity. Although iSPN activation reduces baseline and peak activity in M1 evoked by cues and cued lever presses, activation of dSPNs has no effect on peak firing rates in these periods. This finding cannot be ascribed to a ceiling effect in M1 firing rates, as the peak-firing rate reached during tone-only (failure) trials is well below maximal, yet still unaffected by activation of dSPNs. An alternative explanation is that dSPNs, or circuit elements downstream of dSPNs, are maximally active during the movement such that optogenetic stimulation of dSPNs has no further effect on M1 activity patterns. Such an explanation would also imply that iSPNs are comparatively less active during these periods than dSPNs, which at first may appear in conflict with results observing movement related activity from both iSPNs and dSPNs in vivo (Cui et al., 2013; Isomura et al., 2013). However, this may be reconciled by the greater sustained activity of dSPN compared to iSPNs reported during movement bouts (Jin et al., 2014). Indeed, in motor cortex, we detect larger effects of iSPN activation before the lever press than during the movement itself, indicating that some iSPN activation may be present during the presses.

Two results reveal that not all cortical neurons are equally sensitive to changes in striatal activity, suggesting specificity in either the thalamic target of the BG or of the sub-cortical to cortical projections. First, although changing iSPN or dSPN activity modulated the vast majority of M1 units, a fraction (8–20% depending on the statistical model, see supplemental methods) was insensitive to the optogenetic manipulations of striatum. Although we cannot rule out that different, unstimulated, regions of striatum could modulate these cells, such an explanation would still indicate that neighboring cortical cells are differentially sensitive to non-neighboring regions of striatum.

Second, while the majority of cortical neurons monotonically decrease firing following iSPN activation and increase back to baseline levels upon cessation of iSPN stimulation, ~30% of neurons are transiently excited for ~500 ms following stimulation of iSPNs. These units subsequently reduce their firing rate despite maintained iSPN stimulation, and rebound strongly upon cessation of iSPN stimulation. We were unable to find an analogous class of units that behaved anomalously to initiation and cessation of dSPN stimulation. It is of particular interest that the transiently excited cells appear in predominantly superficial layers. Thalamocortical axons from the VLo, and that are thus likely modulated by the BG, primarily innervate superficial layers (Kuramoto et al., 2009; McFarland and Haber, 2002) (although see (Constantinople and Bruno, 2013)).

At the population level, differences between the cortical effects of iSPN and dSPN activation were also evident. Whereas a strong correlation was observed between each neuron’s modulation by dSPN activation and its lever press related change in firing rate, no similar correlation was found when iSPNs where activated. This is especially intriguing as the iSPN and dSPN projections target the same neurons in SNr (Smith and Bolam, 1991) and thus it is difficult to explain differential effects on cortex via a common output. This may again reflect the existence of functional subsets within the outputs of the BG that are differentially dependent on iSPN and dSPN activity (Saunders et al., 2015).

Conclusions

The results we report here support many predictions of classical models of BG/cortex interactions such that the BG exert large, push-pull control over motor cortex in behaving mice prior to presentation of a reward-associated cue. However, the classic model fails to account for the effects of striatal manipulations when the animals make spontaneous lever presses, and for the asymmetric effects of direct and indirect pathway activation on cortex. Our results suggest the existence of circuitry, either within nuclei downstream of striatum and between the BG and cortex, which allow differential and non-opposing effects of dSPNs and iSPNs on cortex.

Experimental Procedures

Mice expressing cre in iSPNs or dSPNs were injected with virus expressing ChR2 in striatum (0.9A, 1.7L, 2.8D) and surgically implanted with a headpost. Mice were trained in an operant task to press a lever after a tone (50ms 10kHz). ChR2 stimulation (continuous 1.5–3mW) through an implanted fiber optic occurred during extracellular recordings in motor cortex. Full experimental procedures are in the supplement.

Supplementary Material

Acknowledgements

We thank M. Higley for establishing in vivo recordings; W. Wray, H. Hussein, V. Swantic, T. Haynes, and D. Rothfuss for animal training; and the Sabatini lab, J. Maunsell, and J. Assad for helpful discussions. Supported by NIH F31-MH093026-01A1 (IAO) and R01-NS046579 (BLS).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Albin RL, Young AB, Penney JB. The functional anatomy of basal ganglia disorders. Trends in Neurosciences. 1989;12:366–375. doi: 10.1016/0166-2236(89)90074-x. [DOI] [PubMed] [Google Scholar]
  2. Aldridge JW, Anderson RJ, Murphy JT. The role of the basal ganglia in controlling a movement initiated by a visually presented cue. Brain Research. 1980;192:3–16. doi: 10.1016/0006-8993(80)91003-3. [DOI] [PubMed] [Google Scholar]
  3. Alexander GE, Crutcher MD. Functional architecture of basal ganglia circuits: neural substrates of parallel processing. Trends in Neurosciences. 1990;13:266–271. doi: 10.1016/0166-2236(90)90107-l. [DOI] [PubMed] [Google Scholar]
  4. Bateup HS, Santini E, Shen W, Birnbaum S, Valjent E, Surmeier DJ, Fisone G, Nestler EJ, Greengard P. Distinct subclasses of medium spiny neurons differentially regulate striatal motor behaviors. Proc. Natl. Acad. Sci. U.S.a. 2010;107:14845–14850. doi: 10.1073/pnas.1009874107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Constantinople CM, Bruno RM. Deep cortical layers are activated directly by thalamus. Science. 2013;340:1591–1594. doi: 10.1126/science.1236425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cui G, Jun SB, Jin X, Pham MD, Vogel SS, Lovinger DM, Costa RM. Concurrent activation of striatal direct and indirect pathways during action initiation. Nature. 2013;494:238–242. doi: 10.1038/nature11846. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Da Cunha C, Boschen SL, Gómez-A A, Ross EK, Gibson WSJ, Min H-K, Lee KH, Blaha CD. Toward sophisticated basal ganglia neuromodulation: Review on basal ganglia deep brain stimulation. Neurosci Biobehav Rev. 2015 doi: 10.1016/j.neubiorev.2015.02.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. DeLong MR. Primate models of movement disorders of basal ganglia origin. Trends in Neurosciences. 1990;13:281–285. doi: 10.1016/0166-2236(90)90110-v. [DOI] [PubMed] [Google Scholar]
  9. Deniau JM, Chevalier G. Disinhibition as a basic process in the expression of striatal functions. II. The striato-nigral influence on thalamocortical cells of the ventromedial thalamic nucleus. Brain Research. 1985;334:227–233. doi: 10.1016/0006-8993(85)90214-8. [DOI] [PubMed] [Google Scholar]
  10. Freeze BS, Kravitz AV, Hammack N, Berke JD, Kreitzer AC. Control of basal ganglia output by direct and indirect pathway projection neurons. J Neurosci. 2013;33:18531–18539. doi: 10.1523/JNEUROSCI.1278-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gerfen CR, Engber TM, Mahan LC, Susel Z, Chase TN, Monsma FJ, Sibley DR. D1 and D2 dopamine receptor-regulated gene expression of striatonigral and striatopallidal neurons. Science. 1990;250:1429–1432. doi: 10.1126/science.2147780. [DOI] [PubMed] [Google Scholar]
  12. Graybiel AM, Aosaki T, Flaherty AW, Kimura M. The basal ganglia and adaptive motor control. Science. 1994;265:1826–1831. doi: 10.1126/science.8091209. [DOI] [PubMed] [Google Scholar]
  13. Hikosaka O, Sakamoto M, Usui S. Functional properties of monkey caudate neurons. I. Activities related to saccadic eye movements. Journal of Neurophysiology. 1989;61:780–798. doi: 10.1152/jn.1989.61.4.780. [DOI] [PubMed] [Google Scholar]
  14. Isomura Y, Takekawa T, Harukuni R, Handa T, Aizawa H, Takada M, Fukai T. Reward-modulated motor information in identified striatum neurons. J Neurosci. 2013;33:10209–10220. doi: 10.1523/JNEUROSCI.0381-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Jin X, Tecuapetla F, Costa RM. Basal ganglia subcircuits distinctively encode the parsing and concatenation of action sequences. Nature Publishing Group. 2014;17:423–430. doi: 10.1038/nn.3632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Koralek AC, Jin X, Long JD, Costa RM, Carmena JM. Corticostriatal plasticity is necessary for learning intentional neuroprosthetic skills. Nature. 2012;483:331–335. doi: 10.1038/nature10845. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kravitz AV, Freeze BS, Parker PRL, Kay K, Thwin MT, Deisseroth K, Kreitzer AC. Regulation of parkinsonian motor behaviours by optogenetic control of basal ganglia circuitry. Nature. 2010;466:622–626. doi: 10.1038/nature09159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kuramoto E, Furuta T, Nakamura KC, Unzai T, Hioki H, Kaneko T. Two types of thalamocortical projections from the motor thalamic nuclei of the rat: a single neuron-tracing study using viral vectors. Cerebral Cortex. 2009;19:2065–2077. doi: 10.1093/cercor/bhn231. [DOI] [PubMed] [Google Scholar]
  19. Mahon S, Vautrelle N, Pezard L, Slaght SJ, Deniau J-M, Chouvet G, Charpier S. Distinct patterns of striatal medium spiny neuron activity during the natural sleep-wake cycle. J Neurosci. 2006;26:12587–12595. doi: 10.1523/JNEUROSCI.3987-06.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Marsden CD. Basal ganglia disease. Lancet. 1982;2:1141–1147. doi: 10.1016/s0140-6736(82)92797-0. [DOI] [PubMed] [Google Scholar]
  21. McFarland NR, Haber SN. Thalamic relay nuclei of the basal ganglia form both reciprocal and nonreciprocal cortical connections, linking multiple frontal cortical areas. J Neurosci. 2002;22:8117–8132. doi: 10.1523/JNEUROSCI.22-18-08117.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Mink JW. The basal ganglia: focused selection and inhibition of competing motor programs. Prog Neurobiol. 1996;50:381–425. doi: 10.1016/s0301-0082(96)00042-1. [DOI] [PubMed] [Google Scholar]
  23. Mink JW, Thach WT. Basal ganglia motor control. II. Late pallidal timing relative to movement onset and inconsistent pallidal coding of movement parameters. Journal of Neurophysiology. 1991;65:301–329. doi: 10.1152/jn.1991.65.2.301. [DOI] [PubMed] [Google Scholar]
  24. Saunders A, Oldenburg IA, Berezovskii VK, Johnson CA, Kingery ND, Elliott HL, Xie T, Gerfen CR, Sabatini BL. A direct GABAergic output from the basal ganglia to frontal cortex. Nature. 2015 doi: 10.1038/nature14179. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Smith Y, Bolam JP. Convergence of synaptic inputs from the striatum and the globus pallidus onto identified nigrocollicular cells in the rat: a double anterograde labelling study. Neuroscience. 1991;44:45–73. doi: 10.1016/0306-4522(91)90250-r. [DOI] [PubMed] [Google Scholar]
  26. Spampinato U, Girault JA, Danguir J, Savaki HE, Glowinski J, Besson MJ. Apomorphine and haloperidol effects on striatal 3H-dopamine release in anesthetized, awake restrained and freely moving rats. Brain Research Bulletin. 1986;16:161–166. doi: 10.1016/0361-9230(86)90028-6. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS689080-supplement.pdf (3.8MB, pdf)

RESOURCES