How does “learning” turn into memory”? Input type-dependent synaptic dynamics in the motor cortex during learning.

2022.07.28 PressRelease


Researchers from the National Institute for Physiological Sciences show that two presynaptic neural circuits, cortical and thalamic afferents to the motor cortex, supervise distinct programs of synaptic dynamics to execute learning. This study opens a new perspective on the role of dual presynaptic circuit supervision for synaptic dynamics that may allow the transition in motor learning from top-down, goal-directed practice during learning to an automatic and habitual regulation after learning.

Researchers from the National Institute for Physiological Sciences in Japan identify two types of synaptic dynamics that are important for “motor learning” and “motor memory”.

Okazaki, Japan – When we learn a motor pattern such as a new sport or dance, movements that were not possible at the beginning can be memorized and made easy through practice. It has been widely thought that "motor memory" is an extension of "motor learning". However, this study demonstrates that different neural circuits underlie "learning" and "memory".

It has been known that when mice are trained to perform a specific motor task, new synapses are formed in the primary motor cortex, rearranging neural circuits. However, details such as which neuronal connections contribute to learning and memory has not been well understood. Associate Professor Yoshiyuki Kubota and his research team at the Institute for Physiological Sciences have shown that top-down information from the higher-order motor cortex to the primary motor cortex is important for "motor learning" and that the acquired "motor memories" are then taken over and newly stored in synaptic inputs from the thalamus, the gatekeeper for the cortex. The researchers recently published these findings in Science Advances.

The researchers first trained mice for a forelimb reaching task while performing in vivo two-photon imaging of synaptic dynamics in the primary motor cortex. This was followed by post-hoc identification of the presynaptic sources as cortico-cortical (CC) or thalamo-cortical (TC) neurons, to correlate the synaptic plasticity with synaptic input origins. First, top-down CC afferents formed profuse, small synapses in the motor cortex during initial learning. However, these CC synapses were present only transiently, likely being pruned during the subsequent training days. In contrast, although TC synapse formation was less common than CC synapse formation during the early learning period, new TC synapses, one formed, were mostly maintained and matured.

Furthermore, when CC inputs from the higher-order cortical area to the motor cortex was suppressed during motor learning, improvement of the motor skill was impaired, indicating the dependence of “learning” on the CC circuit. On the other hand, suppression of TC inputs did not directly affect task performance. Rather, suppression of TC inputs in mice that had completed motor learning and had established motor memories significantly reduced the success rate, indicating that the acquired “memory” was stored in a TC, and not CC, circuit.

“The significance of this finding is that it offers the first circuit-based mechanism for motor skill learning as a two-step process with initial and transient cortical synaptic supervision to subsequent and stable subcortical control,” Dr. Kubota says. “This study suggests for the first time that motor learning involves two classes of synapses - those specific to learning and those specific to memory. This challenges the widely held view that single synapses mediate both learning and memory.”

Because the neocortex is organized as a repeating lattice presumably with common principles, the mechanism the researchers report in this study for motor learning may eventually be found to apply to other cortical areas and associated learning behaviors.

Release Source

Presynaptic Supervision of Cortical Spine Dynamics in Motor Learning.
Jaerin Sohn, Mototaka Suzuki, Mohammed Youssef, Sayuri Hatada, Matthew E Larkum, Yasuo Kawaguchi, Yoshiyuki Kubota.
Science Advances.
DOI: 10.1126/sciadv.abm0531

Relevant Department

Relevant Researchers