The primary motor cortex (M1) is an essential site for movement execution and motor learning. Within M1, neurons encode movement goals and movement kinematics (Georgopoulos et al., 1992; Moran and Schwartz, 1999; Peters et al., 2014). Intriguingly, neurons in M1 have also been reported to show reward-related activity. In vivo recording studies performed in nonhuman primates found neurons in M1 that encode reward anticipation, reward delivery, and mismatches between the two (Marsh et al., 2015; Ramakrishnan et al., 2017; Ramkumar et al., 2016). In human subjects, reward has also been shown to modulate M1 activity, likely through an inhibitory circuit-dependent mechanism (Thabit et al., 2011). However, it remains unclear how reward-related responses are represented in M1, and if the representation changes with associative learning.

It was recently shown that in well-trained mice performing a skilled reaching task, a subset of layer 2/3 (L2/3) pyramidal neurons (PNs) in M1 specifically report successful, but not failed, reach-and-grasp movements. In contrast, a different subset of PNs report only failed reach-and-grasp movements (Levy et al., 2020). Since the ability to use past experience to learn action–outcome associations is critical to survival, encoding the outcome in M1 may be an important part of motor skill learning. It is widely accepted that associative learning using reinforcement can accelerate and enhance learning (Abe et al., 2011; Nikooyan and Ahmed, 2015). In the case of motor learning, studies have demonstrated that positive feedback (reward) facilitates motor memory retention and negative feedback (punishment) speeds up the learning process (Galea et al., 2015). One hypothesis is that during learning, reward signals in the brain, together with neuromodulators and synaptic plasticity, are involved in potentiating and optimizing the neural circuitry in M1 that underlies the rewarded movement. Implementing such a learning process would necessitate the interplay between different cell types within the local microcircuitry (Richards et al., 2019).

M1, like other cortical areas, is densely packed with PNs and diverse inhibitory interneuron (IN) types and is wired in a delicately balanced and intricate circuit. Different IN subtypes have been shown to have distinct gene expression profiles, electrophysiological properties, and connectivity motifs (Fishell and Rudy, 2011; Markram et al., 2004). Somatostatin-, parvalbumin-, and vasoactive intestinal peptide- expressing inhibitory neurons (SOM-INs, PV-INs, and VIP-INs, respectively) are three major nonoverlapping subtypes of GABAergic neurons that broadly form a common microcircuit motif in the cortex. Some studies have demonstrated that SOM-INs preferentially target distal dendrites of PNs to filter synaptic inputs, fast-spiking PV-INs preferentially target perisomatic regions of PNs enabling strong inhibition of spiking, and VIP-INs regulate local microcircuits by controlling other local INs (Pfeffer et al., 2013). Due to their diverse properties and strategic connectivity motifs, these INs exert fine control over local network activity and provide a potential mechanism for how the brain processes reward signals and ultimately uses this information to optimize neural activity related to learned motor skills.

Multiple studies using in vivo opto-recordings in the primary visual cortex have shown that visual orientation selectivity in PNs is modulated and sharpened by PV- and SOM-INs (Atallah et al., 2012; Lee et al., 2012; Wilson et al., 2012). In the primary auditory cortex, PV- and SOM-INs exert analogous control over PN frequency tuning (Seybold et al., 2015). Moreover, in the auditory cortex, prefrontal cortex, and basolateral amygdala, reinforcement signals such as reward and punishment have been shown to recruit VIP-INs, which in turn, inhibit SOM- and PV-INs (Krabbe et al., 2019; Pi et al., 2013). The subtype-specific roles of these INs have long been elusive, but a complex picture is emerging where INs are not only responsible for maintaining a delicate balance of excitation and inhibition, but are also actively involved in processing activity in the cortex (Lee et al., 2020; Wood et al., 2017).

Here, we employed chronic in vivo two-photon imaging, combined with a head-fixed classical conditioning task, to monitor the activity of the same population of PNs, PV-INs, SOM-INs, or VIP-INs before and after associative learning to investigate whether and how conditioned cue stimulus (CS) and unconditioned reward are represented among different neuronal cell types in M1. Our results demonstrate that all four major cell types in M1 show distinct responses to CS and reward, and their response properties undergo cell-type-specific modifications after associative learning. Notably, PV-INs and VIP-INs exhibited stimulus-specific modifications, in which PV-INs became more reliably responsive to the CS but not to the reward, whereas VIP-INs became more reliable to the reward but not to the CS. PNs initially showed robust responses to novel reward but became habituated to it after associative learning. Lastly, SOM-IN responses emerged with learning and responded more reliably to both the CS and reward. Taken together, these results show that cue- and reward-related signals are preferentially represented among major neuronal cell types in M1, and they undergo cell-type-specific modifications during associative learning, indicating they may have distinct roles in integrating reinforcement signals to promote circuit reorganization in M1 during motor skill learning.

To understand how reward-associated signals are represented within the local microcircuitry in M1 before and after associative learning, we established a head-fixed auditory cued reward conditioning task, which allowed us to combine the task with in vivo two-photon Ca²⁺ imaging to examine the response properties of different neuronal cell-type populations in awake and behaving mice (Figure 1A). In this task, water-restricted mice were exposed to a conditioned stimulus (CS; auditory tone, 1-s duration), followed by a 1.5-s delay and then the delivery of the unconditioned stimulus (US; water reward, ~10 µl). Mice were trained for ~30–35 trials/session (1 session/day for 7 days) with randomly varied intertrial intervals (ITIs) between 60 and 120 s (Figure 1B). Since M1 is known to be involved in movement initiation and motor skill learning, we chose to use a simple classical conditioning task with just an auditory tone paired with reward and omitted any additional training where mice would be required to learn a new movement. The rationale for this is that many neuronal cell types, including PNs, PV-, and SOM-INs, have been shown to undergo modifications when mice acquire new movements (Chen et al., 2015; Cichon and Gan, 2015; Donato et al., 2013; Xu et al., 2009). However, since licking is an innate movement that does not induce plastic changes in adult mice (Chen et al., 2015; Komiyama et al., 2010; Peters et al., 2014), we can reliably attribute changes in neuronal activity over the course of the task to associative learning, rather than motor learning.

Figure 1

Download asset Open asset

Mice learned to associate the CS with the reward after 7 days, shown by an increase in anticipatory lick rate, a conditioned response, following the cue onset on day 7 compared to day 1 (Figure 1C). On a trial-by-trial basis, anticipatory lick rate did not change significantly within a single session on both days 1 and 7, implying limited within-session improvements (Figure 1D). To ensure the increase in lick rate was specific to the CS, we compared the mean lick rate during the CS and reward period to the lick rate during self-initiated spontaneous lick bouts in the ITI (in the absence of the CS or reward). To be consistent with the 2.5-s analysis window for CS responses, we analyzed the first 2.5 s of ITI lick bouts and 2.5 s following reward delivery. On day 1, the mean lick rate during ITI lick bouts (1.47 ± 0.17/s) and the CS (1.37 ± 0.17/s) were similar, while the lick rate following reward delivery was significantly higher (7.16 ± 0.48/s). In contrast, on day 7 following associative learning, the lick rate during the CS period (3.42 ± 0.37/s) was significantly higher than during ITI lick bouts (1.71 ± 0.2/s), demonstrating that the mice effectively learned the CS–reward association by day 7 (Figure 1E, F).

To investigate the activity of different neuronal cell types during this task, we used in vivo two-photon Ca²⁺ imaging of different cell-type populations. To target PNs in M1, we injected an adenoassociated virus (AAV) carrying a Ca²⁺ indicator (GCaMP6f) driven by the CaMKII promoter (AAV1.CaMKIIa.GCaMP6f) into M1 of wild-type B6129S mice. After 3–5 weeks, we recorded the activity of hundreds of L2/3 PNs using two-photon microscopy in awake mice while they underwent the head-fixed conditioning task, and we tracked the same population of neurons on days 1 and 7 (Figure 2A). We identified all the active neurons within a session, irrespective of the behavioral task (see Methods), and sorted neurons by the timing of their peak activity relative to the CS onset. It was apparent that there were subpopulations of neurons more responsive to CS, reward, or both (Figure 2B, C). We also repeated the experiments to examine if the major IN subtypes in M1 also respond to the CS and reward during the conditioning task. To do this, we injected AAV-Syn-Flex-GCaMP6f in PV-Cre, SOM-Cre, or VIP-Cre transgenic mice to selectively express GCaMP6f in PV-INs, SOM-INs, or VIP-INs, respectively, and then performed in vivo two-photon Ca²⁺ imaging to monitor the response properties of the same population of INs on days 1 and 7 after associating learning (Figure 2—figure supplements 1 and 2). We compared the mean percentage of active cells within the entire session to ensure all cell types had a similar proportion of active cells (irrespective of the behavioral task) on days 1 and 7 (Figure 2D).

Figure 2 with 2 supplements see all

Download asset Open asset

To examine task-related activity in each cell type, we first compared the mean percent of active cells during the CS and reward to a null distribution made by randomly sampling the session irrespective of the behavioral task, and then calculating the mean percentage of active neurons during the sampled period. By repeating this 1000 times for each cell type on days 1 and 7, we created a distribution of the percentage of active neurons that were present at baseline levels or by chance. Surprisingly, we found that only PV-IN and VIP-IN cell types had a percent of CS- and reward-responsive cells that were significantly greater than chance level on both days 1 and 7 (PV-IN CS: day 1: 15.26% ± 2.11%, day 7: 24.09% ± 2.98%; PV-IN reward: day 1: 20.17% ± 3.27%, day 7: 27.47% ± 3.66%; VIP-IN CS: day 1: 11.29% ± 3.23%, day 7: 16.59% ± 2.01%, VIP-IN reward: day 1: 18.65% ± 5.91%, day 7: 26.3% ± 6.04%; Figure 2F, G). PN responses to the CS were not different from the null distribution on both days 1 and 7 (day 1: 18.42% ± 1.05%, day 7: 16.41% ± 1.33%; Figure 2E); in contrast, PN responses to the reward were significantly higher on day 1 but significantly lower than the null distribution on day 7 (day 1: 23.95% ± 2.49%, day 7: 12.12% ± 0.95%; Figure 2E). Lastly, SOM-INs showed significant responses to the CS and reward only on day 7 following associative learning, while on day 1, they demonstrated no response to the CS and a modest response to the reward (CS: day 1: 5.53% ± 2.7%, day 7: 13.56% ± 3.17%; Reward: day 1: 9% ± 3.66%, day 7: 12.15% ± 3.93%; Figure 2H). Based on these findings, we decided to only examine the sessions where the percent of active cells were significantly greater than the null distribution in our subsequent analysis, as a nonsignificant percent of active cells during the stimulus period cannot be readily distinguished from nontask-related baseline noise.

We began our analysis on PV-INs and VIP-INs because they both showed significant responses to both the CS and reward on days 1 and 7. To understand how their representations of reward and reward-associated cues changed over the course of learning, we first analyzed the tuning of individual cells to unbiasedly identify their response properties. By quantifying the tuning of each cell’s average response during the CS and reward response periods (2.5-s window) using the nonparametric Spearman correlation ρ (see Methods), we observed a wide range of tuning coefficients to the CS and reward, with a small proportion that was strongly positively or negatively tuned to the CS or reward stimulus (tuning coefficient near –1 or 1; Figure 3A–D), consistent with our earlier analyses demonstrating that neurons in M1 show activity associated with the CS or reward during the conditioning task. We next examined whether the tuning coefficient within each cell type changed after associative learning by calculating the change in tuning coefficients for each cell between days 1 and 7. Again, to validate our findings, we compared these values to a null distribution of Δρ values obtained by randomly sampling the two sessions (see Method details). The PV-IN population did not show any significant changes in either CS or reward tuning between days 1 and 7 (${\overline{\Delta \rho }}_{\text{CS}}=-0.049\pm 0.046$, ${\overline{\Delta \rho }}_{\text{reward}}=0.014\pm 0.054$; Figure 3E, F), indicating that neither CS- nor reward-related tuning became stronger after associative learning. In contrast, VIP-INs’ CS tuning did not change significantly between days 1 and 7 ${\overline{\Delta \rho }}_{CS}=-0.065\pm 0.048$, but VIP-INs’ reward tuning significantly increased on day 7 (${\overline{\Delta \rho }}_{\text{reward}}=0.161\pm 0.086$; Figure 3G, H), suggesting a strengthening of VIP-IN responsivity to reward following associative learning.

Figure 3

Download asset Open asset

Although the tuning properties can reveal changes in task-related responsivity, this analysis is limited in identifying changes at the trial-by-trial level. When we assessed population activity following the CS onset (Figure 4A), it was apparent that a group of PV-INs and VIP-INs were responsive to CS on both days 1 and 7 (Figures 4B and 5B). Hence, by identifying and tracking the same neurons from days 1 to 7, we were able to ask if there was (1) an increase in the number of neurons being recruited as CS or reward responsive during associative learning or (2) a change in the trial-by-trial reliability of CS and reward responses. When we compared the mean percent of CS-responsive neurons on days 1 and 7, we found that the average percent of CS-responsive PV-INs during a trial increased significantly by day 7 (day 1: 15.26% ± 2.11%, day 7: 24.09% ± 2.98%; Figure 4C), while the percent of CS-responsive VIP-INs did not change (day 1: 11.29% ± 3.23%, day 7: 16.59% ± 2.01%; Figure 4D), demonstrating that more PV-INs became responsive to the CS after associative learning. We then assessed the reliability of the responses, defined as the percent of trials within a session where a neuron was responsive to the CS. This measure quantifies how consistently a neuron responded to the CS within a session. We first plotted the cumulative distribution function of reliabilities among all PV-INs and VIP-INs. We observed that PV-INs, as a population, were significantly more reliable in their CS responses than VIP-INs on day 1 (Figure 4E). When we sorted neurons based on their day 1 reliability values and followed them to day 7, we observed that many of the PV-INs that initially had Low Reliability to CS became more responsive on day 7 (Figure 4H); therefore, we grouped neurons into ‘High Reliability’ if they were among the top 50th percentile, while neurons in the lower 50th percentile were deemed ‘Low Reliability’. We found that PV-INs that began as highly reliable maintained their reliability to the CS (day 1: 29.8% ± 1.51%, day 7: 33.87% ± 4.72%), while PV-INs that began as Low Reliability became significantly more reliable (8.47% ± 0.46%, day 7: 18.99% ± 3.76%; Figure 4F). In contrast, the reliability of both High and Low VIP-INs did not change (High Reliability: day 1: 26.55% ± 2.62%, day 7: 25.93% ± 3.81%; Low Reliability: day 1: 6.32% ± 0.76%, day 7: 14.24% ± 2.6%; Figure 4G). We then followed individual Low Reliability PV-INs and calculated the change in reliability to the CS ($∆{reliability}_{CS})$ from days 1 to 7. As a control, we randomly sampled the day 7 session irrespective of the behavioral task and calculated a reliability value. We then subtracted that value from the actual day 1 CS reliability to generate a randomized change in reliability ($∆{reliability}_{random})$ for each Low Reliability neuron (see Methods). When we compared the two distributions, we found that among the Low Reliability PV-INs, $∆{reliability}_{CS}$ was significantly greater than the $∆{reliability}_{random}$ control (Figure 4I). Lastly, we examined if the onset of neuronal activity after the CS changed after associative learning, and if any neurons that were previously responsive to the CS became responsive to the reward only. We did not observe a change in the onset of neuronal activity following CS (Figure 4—figure supplement 1A, see Methods). Furthermore, 97% of the CS-responsive PV-INs from day 1 still showed responsivity to CS on day 7 (Figure 4—figure supplement 1C, D). Altogether, these results show that as a population, more PV-INs became responsive to the CS, and this is mainly due to Low Reliability PV-INs that became more reliable following associative learning.

Figure 4 with 2 supplements see all

Download asset Open asset

Figure 5

Download asset Open asset

To demonstrate that changes in PV-INs’ representation of the CS resulted from associative learning, we conducted additional control experiments to examine their responses to tone when mice received no rewards or nonpaired randomly timed rewards. In the first experiment, water-restricted PV-Cre mice were exposed to the same auditory tone used as the CS (1-s duration) but all water rewards were omitted (randomly varied ITI between 60 and 120 s; ~30 trials/session; 1 session/day for 7 days). We imaged PV-IN activity on both days 1 and 7 and assessed their population responses following the tone onset (Figure 4—figure supplement 2A). We first compared the mean percent of active cells during tone to a null distribution made by randomly sampling the session irrespective of the behavioral task (as in Figure 2). Surprisingly, PV-INs did not show significant tone-responsive cells compared to the chance level on either day 1 or 7 (Figure 4—figure supplement 2B); hence, the average percent of tone-responsive PV-INs per trial also did not increase from days 1 to 7 (Figure 4—figure supplement 2C). Next, in a separate cohort of mice, we exposed water-restricted PV-Cre mice to tone, followed by a ‘nonpaired’ water reward that was given at randomly varied time intervals (40–80 s). Mice were also trained for ~30 trials/session (1 session/day for 7 days) with a randomly varied ITI between 15 and 25 s (Figure 4—figure supplement 2D). We found PV-INs were significantly responsive to the tone stimuli on day 1, similar to what we observed earlier in the CS–reward task (Figure 2F). Interestingly, by day 7, PV-INs no longer responded to the tone stimulus (Figure 4—figure supplement 2G). We next examined if mice that received the tone stimulus with nonpaired water reward learned to associate the two after 7 days. We found the animals did not learn the association, as their conditioned response (tone-evoked anticipatory licking) did not increase at day 7 (Figure 4—figure supplement 2E, F). Unlike the mice that learned the association in the CS–reward task, we did not observe a change in the mean percent of tone-responsive PV-INs from days 1 to 7 in the nonpaired paradigm, and the reliability of Low Reliability PV-INs to tone also did not change (Figure 4—figure supplement 2H, I). Together, these results suggest that PV-INs in M1 do not respond to auditory tone in general, but instead only respond to the tone when the animal actively associates it with reward. Moreover, the changes among PV-INs to the CS tone from days 1 to 7 are specific to associative learning.

We next assessed reward responses among PV-INs and VIP-INs in the same manner but now looked for responses within 2.5 s of the reward delivery time (Figures 4B and 5A, B). We tracked the same neurons from days 1 to 7 and compared the mean percent of reward-responsive neurons. PV-INs and VIP-INs did not show a significant change in the percent of responsive cells per trial (PV-IN: day 1: 20.17% ± 3.27%, day 7: 27.47% ± 3.66%; VIP-IN: day 1: 18.65% ± 5.91%, day 7: 26.3% ± 6.04%; Figure 5C, D). When we examined the cumulative distribution of reliabilities for reward responses between the two cell types, VIP-INs as a population were significantly more reliable than PV-INs on day 1 (Figure 5E). By dividing the cells into High and Low Reliability groups, we found the High Reliability VIP-INs maintained their reliability (VIP-IN High Reliability: day 1: 38.79% ± 2.71%, day 7: 35.88% ± 4.32%), and the Low Reliability VIP-INs became significantly more reliable on day 7 (VIP-IN Low Reliability: day 1: 10.25% ± 1.67%, day 7: 24.06% ± 4.87%; Figure 5G, H). In contrast, both High and Low Reliability PV-INs maintained their reliability to reward (High Reliability: day 1: 35.59% ± 2.81%, day 7: 41.94% ± 5.7%; Low Reliability: day 1: 10.58% ± 0.74%, day 7: 19.59% ± 3.08%; Figure 5F). We also followed individual Low Reliability VIP-INs and calculated the change in reliability to reward (${∆reliability}_{reward}$) from days 1 to 7. As a control, we randomly sampled the day 7 session irrespective of the behavioral task and calculated a random ${∆reliability}_{random}$ for each neuron as described above. When we compared the two distributions, the ${∆reliability}_{reward}$ was significantly greater than the ${∆reliability}_{random}$ , demonstrating that Low Reliability VIP-INs became more reliably responsive to reward (Figure 5I). We also examined if the onset of neuronal activity after reward consumption changed after associative learning, and if neurons that were previously responsive to reward became responsive to the CS only. We did not observe a change in the onset of neuronal activity following reward (Figure 4—figure supplement 1B). Furthermore, 95% of the VIP-INs from day 1 still showed responsivity to reward on day 7 (Figure 4—figure supplement 1E, F). Lastly, to demonstrate that changes in VIP-INs’ representation of reward resulted from associative learning, we examined both PV-IN and VIP-IN responses to reward in mice that were exposed to the nonpaired behavioral paradigm (tone+ randomly timed water rewards; Figure 4—figure supplement 2D–F). Consistent to what we observed in the CS–reward paradigm, both PV-INs and VIP-INs consistently showed higher mean percent of active cells during reward (2.5 s from the first lick after reward delivery) compared to the chance level on both days 1 and 7 (Figure 4—figure supplement 2J–M). However, because these mice did not learn the association between the auditory tone and randomly delivered water reward, we did not see an increase in the reliability of the Low Reliability VIP-INs from days 1 to 7 (Figure 4—figure supplement 2O), or a change in the percent of responsive cells in either PV-INs or VIP-INs (Figure 4—figure supplement 2K–N). Altogether, we found that during associative learning, while the proportion of reward-responsive VIP-INs during a given trial did not change, a subset of VIP-INs that were largely unresponsive to reward on day 1 became more reliably responsive on day 7.

Although PV-INs and VIP-INs were the only cell types that were significantly responsive to both CS and reward on both days 1 and 7, PNs and SOM-INs also had significant responses to specific stimuli on certain days. While PNs did not show significant CS responses when compared to baseline, their reward responses on day 1 were significantly above the null distribution, and they became significantly lower than the null distribution on day 7 (Figure 2E). This result is in line with the change in tuning coefficient (${∆\rho }_{reward}$), which showed a significant decrease in reward tuning between days 1 and 7 (${\overline{\Delta \rho }}_{\text{reward}}=-0.141\pm 0.067$; Figure 6A). Moreover, the cumulative distribution function of PN reliability also shifted significantly to lower reliabilities on day 7 compared to day 1 (Figure 6B). These results indicate that PNs initially responded to novel reward; however, they habituated to the reward following associative learning.

Figure 6

Download asset Open asset

SOM-INs initially had no response to the CS on day 1, but their responses became significant on day 7 (Figure 2H). The change in CS tuning coefficient (${∆\rho }_{tone}$) was not significant (${∆\rho }_{tone}=.059\mathrm{}\pm \mathrm{}0.031$, Figure 6C), suggesting their responsivity did not change with learning. Interestingly, when we assessed the cumulative distribution of CS response reliability on days 1 and 7, the cumulative distribution function shifted to significantly higher reliability values on day 7 (Figure 6C, D). Notably, by day 7, there was a visible reduction in the number of SOM-INs that had 0% reliability to CS on day 1, indicating they were completely unresponsive to the CS on day 1 but not on day 7. Finally, SOM-INs showed modest but significant responses to reward on days 1 and 7 (Figure 2H). When we assessed the reward tuning among the SOM-IN population, ${∆\rho }_{reward}$ did not show a significant change between days 1 and 7 (${∆\rho }_{reward}=.$ 0.017 ± 0.040, Figure 6E). However, SOM-IN reliabilities also shifted to higher values on day 7 (Figure 6F). Altogether, these results suggest that in naive mice, SOM-INs were unresponsive to the CS and modestly responsive to the novel reward; however, following associative learning, SOM-INs became more reliably responsive to both the CS and reward.

Lastly, reward consumption requires innate tongue movements during licking, and since microstimulation of M1 in mice has been shown to evoke tongue and jaw movements (Komiyama et al., 2010), it is crucial to distinguish whether the observed CS and reward responses resulted from task-related stimuli or if the activity is simply associated with licking movements. We demonstrated earlier that head-fixed mice learned the CS–reward association by displaying the conditioned response (anticipatory licking) following the CS on day 7 (Figure 1). To address this potential confound, we identified all the self-initiated licking bouts during ITIs, when no reward was present (Figure 7A–C; Figure 2—figure supplement 1). We first assessed all the significantly active cells in each cell type (identified in Figure 2D) during the first lick bout of each ITI on days 1 and 7. We observed that in each cell type, the majority of the neurons were nonlick neurons on both days 1 and 7 (Figure 7—figure supplement 1A). We then tracked individual lick and nonlick neurons and examined if they shifted their responses after associative learning. We found that most of the neurons maintained the same responses, and nonlick neurons were still the majority in all cell types (Figure 7—figure supplement 1B). Next, we examined whether the lick neurons also showed mixed responses to CS, reward, or CS+ reward. Indeed, lick neurons exhibited mixed responses to CS, reward, or CS+ reward (Figure 7—figure supplement 1C). We then further divided them into three categories – ‘CS cells’, ‘reward cells’, and ‘CS+ reward cells’ and compared the percentage of neurons in each category between days 1 and 7; we did not observe a significant difference (Figure 7—figure supplement 1D). Lastly, at the population level, we examined the response reliability index of all the active neurons during all ITI lick bouts and compared them to the response reliability index for the CS and reward. On both day 1 (when there was minimal anticipatory licking during the CS) and day 7 (when mice showed anticipatory licking), all cell types exhibited lower reliability index values for the ITI lick bouts compared to the CS and reward, indicating that the increase in task-related responses following water rewards was specific to the reward stimulus, and not licking movements (Figure 7D–J). Together, these results suggest that the cell-type-specific modifications observed between days 1 and 7 were not caused by licking movements.

Figure 7 with 1 supplement see all

Download asset Open asset

M1 is known to be involved in motor initiation, movement kinematics, and motor learning. Recent studies have demonstrated reward-related activity in M1 using in vivo electrophysiological recordings in nonhuman primates (Marsh et al., 2015; Ramakrishnan et al., 2017; Ramkumar et al., 2016) and transcranial magnetic stimulation in human subjects (Thabit et al., 2011). However, whether CS- and reward-associated signals are represented among different neuronal cell types within the microcircuit in M1 is still unclear. Using chronic two-photon Ca²⁺ imaging, combined with transgenic mouse lines and viral strategies to target different neuronal cell types, we demonstrated that during a conditioning task, all major cell types in M1 responded to either the CS, the reward, or both. Most notably, each cell type underwent distinct modifications after association learning. By tracking the same population of neurons, we revealed that the CS-responding population increased among PV-INs and individual cells responded more reliably to the CS following associative learning. On the contrary, VIP-INs became more reliable in response to reward. When mice underwent control behavioral paradigms where tone was not paired with reward and no associative learning occurred, PV-INs and VIP-INs did not undergo these changes. Additionally, PNs had a drastically reduced response to reward, while SOM-INs became more reliable to both the CS and the reward. Our findings suggest that each cell type has a distinct role in processing information related to the cue–reward association in M1, and they may work together to provide the reinforcement signals in M1 that are important for motor skill learning.

Previous studies in trained rhesus monkeys performing a joystick center-out task have shown a widespread representation of reward anticipation and reward-related activity among cortical neurons in M1 (Ramakrishnan et al., 2017). Consistent with earlier work, we also observed reward-related activity in all four major cell types in M1, even in naive mice on day 1 when they were first exposed to the CS and reward. It has been reported that in sensory cortices, repeated passive exposure to a sensory stimulus leads to a long-lasting reduction in PN responsivity, but when animals are engaged in learning, PNs maintain their responsivity to the repeated stimulus (Kato et al., 2015; Makino and Komiyama, 2015). However, we found in M1, when water-restricted mice were engaged in a conditioning task to learn the association between the CS and water reward, PNs still showed a drastic habituation to the reward stimulus. A recent study that imaged neuronal activity in expert mice performing a head-fixed pellet reaching task demonstrated that L2/3 PNs in M1 are involved in encoding movement outcome (success vs. failure) but not the appetitive outcome (reward vs. no reward). However, the authors did not image the mice at the naive stage (Levy et al., 2020). Hence, one possibility is that L2/3 PNs in M1 encode reward signals during the naive stage, but after associative learning, they habituate and become unresponsive to the reward stimulus. In addition, in the sensory cortices, the flexibility to either respond to or ignore sensory stimuli is based on the stimulus’ behavioral relevance and is gated by local SOM-INs (Kato et al., 2015; Makino and Komiyama, 2015; Poort et al., 2021). In line with these findings, we found that SOM-INs became more reliably responsive to both the CS and the reward with associative learning. We also observed stimulus-specific increases in the reliability of PV-INs’ response to the CS and VIP-INs’ response to the reward after associative learning, and these changes to tone and reward were not observed in the absence of associative learning. When mice were exposed to tone alone, PV-INs did not show significant responses to tone on either day 1 or 7, while mice exposed to a nonpaired tone and reward task showed significant responses to tone on day 1, but not on day 7. This suggests that in M1, PV-INs only respond to behaviorally relevant cues such as those that predict reward. Finally, while VIP-INs remained responsive to reward in the nonpaired paradigm, VIP-INs did not show changes in reliability in the absence of associative learning. Together, our results suggest that different IN subtypes may have distinct roles in processing CS- and reward-related information in M1 during motivated associative learning.

One hypothesis is that PV-INs are recruited by the CS to control the behavioral responses (anticipatory licking) during reward anticipation since PV-INs are known to regulate PN firing through both feedforward and feedback inhibition (Fishell and Rudy, 2011; Xu and Callaway, 2009; Xue et al., 2014). Similar observations have been reported in the striatum, in which optogenetic activation or suppression of PV-INs during a similar conditioning task impaired anticipatory licking, demonstrating the importance of PV-INs in the expression of conditioned responses (Lee et al., 2017). Likewise, PV-INs in the basolateral amygdala, are also recruited during the CS and subsequently inhibited during the US in an auditory fear conditioning task. Optogenetic activation of PV-INs during the CS increased conditioned freezing behavior while PV-IN suppression reduced freezing, indicating bidirectional control of the conditioned response (Wolff et al., 2014). Our results demonstrate that in a naive animal, a subset of PV-INs in M1 are responsive to the CS only when rewards are present, and more PV-INs are recruited by the CS if the animal learns that the CS predicts reward. This suggests that in M1, PV-IN responses to the CS are not purely sensory, but rather, they may play an important role in controlling the behavioral responses to the CS.

VIP-INs, on the other hand, were significantly less reliable in responding to the CS compared to PV-INs, and their responses to the CS remained low. However, VIP-INs’ responses to the reward were more reliable than those of PV-INs, and they became more closely tuned and reliably responsive to the reward with learning. Due to the disinhibitory position of VIP-INs in the microcircuit, activation of VIP-INs can lead to widespread increases in local excitability and contribute to regulating cortical gain (Fu et al., 2014; Jackson et al., 2016; Pfeffer et al., 2013). Furthermore, a growing body of evidence suggests a general principle across brain regions, in which VIP-INs receive long-range inputs (Duan et al., 2020; Krabbe et al., 2019; Turi et al., 2019; Zhang et al., 2014; Gasselin et al., 2021), respond to reinforcement signals (Krabbe et al., 2019; Pi et al., 2013), and play an important role in goal-oriented learning (Krabbe et al., 2019; Turi et al., 2019). Taken together, our results suggest that during CS–reward conditioning, PV-INs in M1 encode the CS association, and may regulate local circuit activity related to reward anticipation, whereas VIP-INs act as a context-dependent switch following the reward delivery (Muñoz et al., 2017; Turi et al., 2019) to instruct and disinhibit local PNs to enable learning-induced plastic changes critical for the acquisition of new movements. An interesting point to note is that since PV-INs are only responsive to tone when it is paired with reward, and neither PV-INs nor VIP-INs undergo plastic changes in the absence of associative learning, M1 is unlikely to be a primary site for learning reward predictions. We hypothesize that other brain regions are responsible for learning relevant CS–reward associations while filtering out behaviorally irrelevant stimuli, and these regions subsequently send long-range inputs to M1 to instruct motor responses to the CS and the reward. In summary, this study provides insight on how different IN subtypes in M1 integrate incoming inputs from various brain regions and orchestrate local circuit plasticity. Future work will be important to identify the origin of these putative long-range inputs to different cell types in M1.

Codes to reproduce the analysis for figures 1-2 and 4-7 are available at https://github.com/clee162/Analysis-of-Cell-type-Specific-Responses-to-Associative-Learning-in-M1. Codes to reproduce the analysis and figure 3 are available at https://github.com/nauralcodinglab/interneuron-reward. Data can be found on Dryad at https://doi.org/10.5061/dryad.q573n5tjj.

1. 1. Chen SX

   (2022) Dryad Digital Repository

   Data from: Cell-type specific responses to associative learning in the primary motor cortex.

   https://doi.org/10.5061/dryad.q573n5tjj

1. 1. Abe M
   2. Schambra H
   3. Wassermann EM
   4. Luckenbaugh D
   5. Schweighofer N
   6. Cohen LG

   (2011) Reward improves long-term retention of a motor memory through induction of offline memory gains

   Current Biology 21:557–562.

   https://doi.org/10.1016/j.cub.2011.02.030

   • PubMed
   • Google Scholar
2. 1. Atallah BV
   2. Bruns W
   3. Carandini M
   4. Scanziani M

   (2012) Parvalbumin-expressing interneurons linearly transform cortical responses to visual stimuli

   Neuron 73:159–170.

   https://doi.org/10.1016/j.neuron.2011.12.013

   • PubMed
   • Google Scholar
3. 1. Chen SX
   2. Kim AN
   3. Peters AJ
   4. Komiyama T

   (2015) Subtype-specific plasticity of inhibitory circuits in motor cortex during motor learning

   Nature Neuroscience 18:1109–1115.

   https://doi.org/10.1038/nn.4049

   • PubMed
   • Google Scholar
4. 1. Chu MW
   2. Li WL
   3. Komiyama T

   (2016) Balancing the Robustness and Efficiency of Odor Representations during Learning

   Neuron 92:174–186.

   https://doi.org/10.1016/j.neuron.2016.09.004

   • PubMed
   • Google Scholar
5. 1. Cichon J
   2. Gan WB

   (2015) Branch-specific dendritic Ca(2+) spikes cause persistent synaptic plasticity

   Nature 520:180–185.

   https://doi.org/10.1038/nature14251

   • PubMed
   • Google Scholar
6. 1. Donato F
   2. Rompani SB
   3. Caroni P

   (2013) Parvalbumin-expressing basket-cell network plasticity induced by experience regulates adult learning

   Nature 504:272–276.

   https://doi.org/10.1038/nature12866

   • PubMed
   • Google Scholar
7. 1. Driscoll LN
   2. Pettit NL
   3. Minderer M
   4. Chettih SN
   5. Harvey CD

   (2017) Dynamic Reorganization of Neuronal Activity Patterns in Parietal Cortex

   Cell 170:986–999.

   https://doi.org/10.1016/j.cell.2017.07.021

   • PubMed
   • Google Scholar
8. 1. Duan Z
   2. Li A
   3. Gong H
   4. Li X

   (2020) A Whole-brain Map of Long-range Inputs to GABAergic Interneurons in the Mouse Caudal Forelimb Area

   Neuroscience Bulletin 36:493–505.

   https://doi.org/10.1007/s12264-019-00458-6

   • PubMed
   • Google Scholar
9. 1. Fishell G
   2. Rudy B

   (2011) Mechanisms of inhibition within the telencephalon: “where the wild things are.”

   Annual Review of Neuroscience 34:535–567.

   https://doi.org/10.1146/annurev-neuro-061010-113717

   • PubMed
   • Google Scholar
10. 1. Fu Y
    2. Tucciarone JM
    3. Espinosa JS
    4. Sheng N
    5. Darcy DP
    6. Nicoll RA
    7. Huang ZJ
    8. Stryker MP

    (2014) A cortical circuit for gain control by behavioral state

    Cell 156:1139–1152.

    https://doi.org/10.1016/j.cell.2014.01.050

    • PubMed
    • Google Scholar
11. 1. Galea JM
    2. Mallia E
    3. Rothwell J
    4. Diedrichsen J

    (2015) The dissociable effects of punishment and reward on motor learning

    Nature Neuroscience 18:597–602.

    https://doi.org/10.1038/nn.3956

    • PubMed
    • Google Scholar
12. 1. Gasselin C
    2. Hohl B
    3. Vernet A
    4. Crochet S
    5. Petersen CCH

    (2021) Cell-type-specific nicotinic input disinhibits mouse barrel cortex during active sensing

    Neuron 109:778–787.

    https://doi.org/10.1016/j.neuron.2020.12.018

    • PubMed
    • Google Scholar
13. 1. Georgopoulos AP
    2. Ashe J
    3. Smyrnis N
    4. Taira M

    (1992) The motor cortex and the coding of force

    Science (New York, N.Y.) 256:1692–1695.

    https://doi.org/10.1126/science.256.5064.1692

    • PubMed
    • Google Scholar
14. Software

    1. Harkin EF

    (2022) Calcium activity in M1 during classical conditioning, version 3c30ebc

    GitHub.

    https://github.com/nauralcodinglab/interneuron-reward
15. 1. Harvey CD
    2. Coen P
    3. Tank DW

    (2012) Choice-specific sequences in parietal cortex during a virtual-navigation decision task

    Nature 484:62–68.

    https://doi.org/10.1038/nature10918

    • PubMed
    • Google Scholar
16. 1. Jackson J
    2. Ayzenshtat I
    3. Karnani MM
    4. Yuste R

    (2016) VIP+ interneurons control neocortical activity across brain states

    Journal of Neurophysiology 115:3008–3017.

    https://doi.org/10.1152/jn.01124.2015

    • PubMed
    • Google Scholar
17. 1. Kato HK
    2. Chu MW
    3. Isaacson JS
    4. Komiyama T

    (2012) Dynamic sensory representations in the olfactory bulb: modulation by wakefulness and experience

    Neuron 76:962–975.

    https://doi.org/10.1016/j.neuron.2012.09.037

    • PubMed
    • Google Scholar
18. 1. Kato HK
    2. Gillet SN
    3. Isaacson JS

    (2015) Flexible Sensory Representations in Auditory Cortex Driven by Behavioral Relevance

    Neuron 88:1027–1039.

    https://doi.org/10.1016/j.neuron.2015.10.024

    • PubMed
    • Google Scholar
19. 1. Komiyama T
    2. Sato TR
    3. O’Connor DH
    4. Zhang Y-X
    5. Huber D
    6. Hooks BM
    7. Gabitto M
    8. Svoboda K

    (2010) Learning-related fine-scale specificity imaged in motor cortex circuits of behaving mice

    Nature 464:1182–1186.

    https://doi.org/10.1038/nature08897

    • PubMed
    • Google Scholar
20. 1. Krabbe S
    2. Paradiso E
    3. d’Aquin S
    4. Bitterman Y
    5. Courtin J
    6. Xu C
    7. Yonehara K
    8. Markovic M
    9. Müller C
    10. Eichlisberger T
    11. Gründemann J
    12. Ferraguti F
    13. Lüthi A

    (2019) Adaptive disinhibitory gating by VIP interneurons permits associative learning

    Nature Neuroscience 22:1834–1843.

    https://doi.org/10.1038/s41593-019-0508-y

    • PubMed
    • Google Scholar
21. 1. Lee S-H
    2. Kwan AC
    3. Zhang S
    4. Phoumthipphavong V
    5. Flannery JG
    6. Masmanidis SC
    7. Taniguchi H
    8. Huang ZJ
    9. Zhang F
    10. Boyden ES
    11. Deisseroth K
    12. Dan Y

    (2012) Activation of specific interneurons improves V1 feature selectivity and visual perception

    Nature 488:379–383.

    https://doi.org/10.1038/nature11312

    • PubMed
    • Google Scholar
22. 1. Lee K
    2. Holley SM
    3. Shobe JL
    4. Chong NC
    5. Cepeda C
    6. Levine MS
    7. Masmanidis SC

    (2017) Parvalbumin Interneurons Modulate Striatal Output and Enhance Performance during Associative Learning

    Neuron 93:1451–1463.

    https://doi.org/10.1016/j.neuron.2017.02.033

    • PubMed
    • Google Scholar
23. 1. Lee C
    2. Lavoie A
    3. Liu J
    4. Chen SX
    5. Liu BH

    (2020) Light Up the Brain: The Application of Optogenetics in Cell-Type Specific Dissection of Mouse Brain Circuits

    Frontiers in Neural Circuits 14:18.

    https://doi.org/10.3389/fncir.2020.00018

    • PubMed
    • Google Scholar
24. Software

    1. Lee C

    (2022) Analysis of cell-type-specific responses to associative learning in M1, version 824cf3c

    GitHub.

    https://github.com/clee162/Analysis-of-Cell-type-Specific-Responses-to-Associative-Learning-in-M1
25. 1. Levy S
    2. Lavzin M
    3. Benisty H
    4. Ghanayim A
    5. Dubin U
    6. Achvat S
    7. Brosh Z
    8. Aeed F
    9. Mensh BD
    10. Schiller Y
    11. Meir R
    12. Barak O
    13. Talmon R
    14. Hantman AW
    15. Schiller J

    (2020) Cell-Type-Specific Outcome Representation in the Primary Motor Cortex

    Neuron 107:954–971.

    https://doi.org/10.1016/j.neuron.2020.06.006

    • PubMed
    • Google Scholar
26. 1. Makino H
    2. Komiyama T

    (2015) Learning enhances the relative impact of top-down processing in the visual cortex

    Nature Neuroscience 18:1116–1122.

    https://doi.org/10.1038/nn.4061

    • PubMed
    • Google Scholar
27. 1. Markram H
    2. Toledo-Rodriguez M
    3. Wang Y
    4. Gupta A
    5. Silberberg G
    6. Wu C

    (2004) Interneurons of the neocortical inhibitory system

    Nature Reviews. Neuroscience 5:793–807.

    https://doi.org/10.1038/nrn1519

    • PubMed
    • Google Scholar
28. 1. Marsh BT
    2. Tarigoppula VSA
    3. Chen C
    4. Francis JT

    (2015) Toward an autonomous brain machine interface: integrating sensorimotor reward modulation and reinforcement learning

    The Journal of Neuroscience 35:7374–7387.

    https://doi.org/10.1523/JNEUROSCI.1802-14.2015

    • PubMed
    • Google Scholar
29. 1. Moran DW
    2. Schwartz AB

    (1999) Motor cortical representation of speed and direction during reaching

    Journal of Neurophysiology 82:2676–2692.

    https://doi.org/10.1152/jn.1999.82.5.2676

    • PubMed
    • Google Scholar
30. 1. Muñoz W
    2. Tremblay R
    3. Levenstein D
    4. Rudy B

    (2017) Layer-specific modulation of neocortical dendritic inhibition during active wakefulness

    Science (New York, N.Y.) 355:954–959.

    https://doi.org/10.1126/science.aag2599

    • PubMed
    • Google Scholar
31. 1. Nikooyan AA
    2. Ahmed AA

    (2015) Reward feedback accelerates motor learning

    Journal of Neurophysiology 113:633–646.

    https://doi.org/10.1152/jn.00032.2014

    • PubMed
    • Google Scholar
32. 1. O’Connor DH
    2. Hires SA
    3. Guo ZV
    4. Li N
    5. Yu J
    6. Sun Q-Q
    7. Huber D
    8. Svoboda K

    (2013) Neural coding during active somatosensation revealed using illusory touch

    Nature Neuroscience 16:958–965.

    https://doi.org/10.1038/nn.3419

    • PubMed
    • Google Scholar
33. 1. Peters AJ
    2. Chen SX
    3. Komiyama T

    (2014) Emergence of reproducible spatiotemporal activity during motor learning

    Nature 510:263–267.

    https://doi.org/10.1038/nature13235

    • PubMed
    • Google Scholar
34. 1. Peters AJ
    2. Lee J
    3. Hedrick NG
    4. O’Neil K
    5. Komiyama T

    (2017) Reorganization of corticospinal output during motor learning

    Nature Neuroscience 20:1133–1141.

    https://doi.org/10.1038/nn.4596

    • PubMed
    • Google Scholar
35. 1. Pfeffer CK
    2. Xue M
    3. He M
    4. Huang ZJ
    5. Scanziani M

    (2013) Inhibition of inhibition in visual cortex: the logic of connections between molecularly distinct interneurons

    Nature Neuroscience 16:1068–1076.

    https://doi.org/10.1038/nn.3446

    • PubMed
    • Google Scholar
36. 1. Pi HJ
    2. Hangya B
    3. Kvitsiani D
    4. Sanders JI
    5. Huang ZJ
    6. Kepecs A

    (2013) Cortical interneurons that specialize in disinhibitory control

    Nature 503:521–524.

    https://doi.org/10.1038/nature12676

    • PubMed
    • Google Scholar
37. 1. Poort J
    2. Wilmes KA
    3. Blot A
    4. Chadwick A
    5. Sahani M
    6. Clopath C
    7. Mrsic-Flogel TD
    8. Hofer SB
    9. Khan AG

    (2021) Learning and attention increase visual response selectivity through distinct mechanisms

    Neuron 1:00954–00955.

    https://doi.org/10.1016/j.neuron.2021.11.016

    • PubMed
    • Google Scholar
38. 1. Ramakrishnan A
    2. Byun YW
    3. Rand K
    4. Pedersen CE
    5. Lebedev MA
    6. Nicolelis MAL

    (2017) Cortical neurons multiplex reward-related signals along with sensory and motor information

    PNAS 114:E4841–E4850.

    https://doi.org/10.1073/pnas.1703668114

    • PubMed
    • Google Scholar
39. 1. Ramkumar P
    2. Dekleva B
    3. Cooler S
    4. Miller L
    5. Kording K

    (2016) Premotor and Motor Cortices Encode Reward

    PLOS ONE 11:e0160851.

    https://doi.org/10.1371/journal.pone.0160851

    • PubMed
    • Google Scholar
40. 1. Richards BA
    2. Lillicrap TP
    3. Beaudoin P
    4. Bengio Y
    5. Bogacz R
    6. Christensen A
    7. Clopath C
    8. Costa RP
    9. de Berker A
    10. Ganguli S
    11. Gillon CJ
    12. Hafner D
    13. Kepecs A
    14. Kriegeskorte N
    15. Latham P
    16. Lindsay GW
    17. Miller KD
    18. Naud R
    19. Pack CC
    20. Poirazi P
    21. Roelfsema P
    22. Sacramento J
    23. Saxe A
    24. Scellier B
    25. Schapiro AC
    26. Senn W
    27. Wayne G
    28. Yamins D
    29. Zenke F
    30. Zylberberg J
    31. Therien D
    32. Kording KP

    (2019) A deep learning framework for neuroscience

    Nature Neuroscience 22:1761–1770.

    https://doi.org/10.1038/s41593-019-0520-2

    • PubMed
    • Google Scholar
41. 1. Seybold BA
    2. Phillips EAK
    3. Schreiner CE
    4. Hasenstaub AR

    (2015) Inhibitory Actions Unified by Network Integration

    Neuron 87:1181–1192.

    https://doi.org/10.1016/j.neuron.2015.09.013

    • PubMed
    • Google Scholar
42. 1. Thabit MN
    2. Nakatsuka M
    3. Koganemaru S
    4. Fawi G
    5. Fukuyama H
    6. Mima T

    (2011) Momentary reward induce changes in excitability of primary motor cortex

    Clinical Neurophysiology 122:1764–1770.

    https://doi.org/10.1016/j.clinph.2011.02.021

    • PubMed
    • Google Scholar
43. 1. Thévenaz P
    2. Ruttimann UE
    3. Unser M

    (1998) A pyramid approach to subpixel registration based on intensity

    IEEE Transactions on Image Processing 7:27–41.

    https://doi.org/10.1109/83.650848

    • PubMed
    • Google Scholar
44. 1. Turi GF
    2. Li W-K
    3. Chavlis S
    4. Pandi I
    5. O’Hare J
    6. Priestley JB
    7. Grosmark AD
    8. Liao Z
    9. Ladow M
    10. Zhang JF
    11. Zemelman BV
    12. Poirazi P
    13. Losonczy A

    (2019) Vasoactive Intestinal Polypeptide-Expressing Interneurons in the Hippocampus Support Goal-Oriented Spatial Learning

    Neuron 101:1150–1165.

    https://doi.org/10.1016/j.neuron.2019.01.009

    • PubMed
    • Google Scholar
45. 1. Wilson NR
    2. Runyan CA
    3. Wang FL
    4. Sur M

    (2012) Division and subtraction by distinct cortical inhibitory networks in vivo

    Nature 488:343–348.

    https://doi.org/10.1038/nature11347

    • PubMed
    • Google Scholar
46. 1. Wolff SBE
    2. Gründemann J
    3. Tovote P
    4. Krabbe S
    5. Jacobson GA
    6. Müller C
    7. Herry C
    8. Ehrlich I
    9. Friedrich RW
    10. Letzkus JJ
    11. Lüthi A

    (2014) Amygdala interneuron subtypes control fear learning through disinhibition

    Nature 509:453–458.

    https://doi.org/10.1038/nature13258

    • PubMed
    • Google Scholar
47. 1. Wood KC
    2. Blackwell JM
    3. Geffen MN

    (2017) Cortical inhibitory interneurons control sensory processing

    Current Opinion in Neurobiology 46:200–207.

    https://doi.org/10.1016/j.conb.2017.08.018

    • PubMed
    • Google Scholar
48. 1. Xu X
    2. Callaway EM

    (2009) Laminar specificity of functional input to distinct types of inhibitory cortical neurons

    The Journal of Neuroscience 29:70–85.

    https://doi.org/10.1523/JNEUROSCI.4104-08.2009

    • PubMed
    • Google Scholar
49. 1. Xu T
    2. Yu X
    3. Perlik AJ
    4. Tobin WF
    5. Zweig JA
    6. Tennant K
    7. Jones T
    8. Zuo Y

    (2009) Rapid formation and selective stabilization of synapses for enduring motor memories

    Nature 462:915–919.

    https://doi.org/10.1038/nature08389

    • PubMed
    • Google Scholar
50. 1. Xue M
    2. Atallah BV
    3. Scanziani M

    (2014) Equalizing excitation-inhibition ratios across visual cortical neurons

    Nature 511:596–600.

    https://doi.org/10.1038/nature13321

    • PubMed
    • Google Scholar
51. 1. Zhang S
    2. Xu M
    3. Kamigaki T
    4. Hoang Do JP
    5. Chang WC
    6. Jenvay S
    7. Miyamichi K
    8. Luo L
    9. Dan Y

    (2014) Selective attention: Long-range and local circuits for top-down modulation of visual cortex processing

    Science (New York, N.Y.) 345:660–665.

    https://doi.org/10.1126/science.1254126

    • PubMed
    • Google Scholar

Author details

1. Candice Lee

   Department of Cellular and Molecular Medicine, University of Ottawa, Ottawa, Canada

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Supervision, Validation, Visualization, Writing - original draft, Writing - review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-7043-9367

2. Emerson F Harkin

   Department of Cellular and Molecular Medicine, University of Ottawa, Ottawa, Canada

   Contribution

   Formal analysis, performed the analyses for Figure 3, performed the analyses for Figure 3

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0698-5894

3. Xuming Yin

   Department of Cellular and Molecular Medicine, University of Ottawa, Ottawa, Canada

   Contribution

   Data curation

   Competing interests

   No competing interests declared

4. Richard Naud

   1. Department of Cellular and Molecular Medicine, University of Ottawa, Ottawa, Canada
   2. Department of Physics, STEM Complex, University of Ottawa, Ottawa, Canada
   3. Brain and Mind Research Institute, University of Ottawa, Ottawa, Canada
   4. Center for Neural Dynamics, University of Ottawa, Ottawa, Canada

   Contribution

   Formal analysis, Supervision, performed the analyses for Figure 3, performed the analyses for Figure 3

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-7383-3095

5. Simon Chen

   1. Department of Cellular and Molecular Medicine, University of Ottawa, Ottawa, Canada
   2. Brain and Mind Research Institute, University of Ottawa, Ottawa, Canada
   3. Center for Neural Dynamics, University of Ottawa, Ottawa, Canada

   Contribution

   Conceptualization, Funding acquisition, Supervision, Writing - original draft, Writing - review and editing

   For correspondence

   schen2@uottawa.ca

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7595-7443

• 2,900

  views

• 375

  downloads

• 17

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.