Contributed Talk Sessions | Poster Sessions | All Posters | Search Papers

Poster Session A: Tuesday, August 12, 1:30 – 4:30 pm, de Brug & E‑Hall

Association Learning with Myelin Plasticity

Charles Shvartsman¹, Pawan Sinha¹; ¹Massachusetts Institute of Technology

Presenter: Charles Shvartsman

Myelin wraps axons, resulting in increased conduction speed. For a long time, myelin patterns were thought to be established in development, after which they remained static. However, over the past decade, evidence has accumulated demonstrating that myelination changes throughout life, is activity dependent, and that myelin plasticity plays a role in various forms of learning, ultimately shaping behavior. Crucially, blocking the formation of new myelin leads to impaired learning in various domains. Despite the experimental evidence, there has been relatively little computational work exploring myelin plasticity, with no plausible account to how low-level changes in myelin can lead to high-level changes in behavior. Here, we take a step towards such an account. Specifically, we demonstrate that a model of spiking neurons, employing established cortical motifs, and using a simple and biologically grounded myelin learning rule, can learn associations – a cognitive building block that underlies more complex capacities such as motor sequencing and predictive processing. Crucially, synaptic weights are all equal and remain static throughout our simulations; the functional changes we observe arise solely from changes in conduction delays. Our work provides a proof of principle as to how myelin plasticity may shape neural circuits to qualitatively change behavior.

Topic Area: Memory, Spatial Cognition & Skill Learning

Extended Abstract: Full Text PDF