Repetitive Practice Improves Working Memory, Changes Brain Pathways

A new study from UCLA Health finds that repeated practice not only helps improve skills, but also leads to significant changes in the brain's memory pathways.

The study, published in Nature and conducted in collaboration with The Rockefeller University, sought to reveal how the brain's ability to store and process information, known as working memory, is improved by training.

To test this, the researchers had mice identify and remember a sequence of odors for two weeks. The researchers tracked neural activity in the animals as they performed the task, using a new custom-built microscope to image the cellular activity of up to 73,000 neurons simultaneously throughout the cerebral cortex.

The study found transformations in working memory circuits located in the secondary motor cortex as mice repeated the task over time. When the mice first started learning the task, the memory representations were unstable. However, after repeated practice of the task, memory patterns began to stabilize or "crystallize," said study lead author and UCLA Health neurologist Dr. Payman Golshani.

Effect of optogenetic inhibition on working memory (WM) task performance.
a. Experimental setup.
b. Trial types in the delayed association WM task; licking was assessed during a 3 second choice period, with early and late delay periods marked.
c. Progression of learning over eight sessions, measured by percentage of correct answers.
d. Example of a training session, with licks marked.
e. Effect of photoinhibition on task performance across different epochs (fourth second of the delay period, P = 0.009; fifth second of the delay period, P = 0.005; second odor, P = 0.0004; first second of the choice period, P = 0.0001). Statistical analysis was performed using paired t-tests.
f. Photoinhibition of M2 in the last 2 seconds of the delay period during the first 7 days of training impairs task performance. N = 4 (stGtACR2-expressing mice) and n = 4 (mCherry-expressing mice). P values determined using two-sample t tests for sessions 1–10 were as follows: P1 = 0.8425, P2 = 0.4610, P3 = 0.6904, P4 = 0.0724, P5 = 0.0463, P6 = 0.0146, P7 = 0.0161, P8 = 0.7065, P9 = 0.6530 and P10 = 0.7955. For c, e and f, data are presented as mean ± s.e.m. NS, not significant; *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.
Source: Nature (2024). DOI: 10.1038/s41586-024-07425-w

“If you imagine that each neuron in the brain sounded like a different note, the melody the brain generated when performing a task varied from day to day, but then became more and more refined and similar as the animals continued to practice task,” Golshani said.

These changes provide insight into why performance becomes more accurate and automatic after repeated practice.

“This discovery not only advances our understanding of learning and memory, but also has implications for addressing problems associated with memory impairment,” Golshani said.

The work was carried out by Dr. Arash Bellafard, a UCLA project scientist, in close collaboration with the group of Dr. Alipasha Vaziri at The Rockefeller University.