Repetition Doesn’t Make You Learn Faster, Spacing Does
October 7, 2026
The number of repetitions has zero influence on overall learning when spaced out over a fixed period. This gap in our scientific understanding has limited how we design everything from educational curricula to AI training protocols. The guest on this episode, Prof. Vijay Mohan K. Namboodiri, leads a research lab at UCSF where he investigates the biological implementation of learning algorithms and dopamine signaling.
Think about the last time you tried to master a new skill, like learning a few phrases in a new language or practicing a difficult chord on a guitar. You likely spent hours drilling the same movements or words, convinced that if you just repeated the task one more time, it would finally stick.
Often, despite all that effort, the progress feels slow, and the information disappears by the next morning. It feels like the brain has a limit on how much it can absorb through sheer force of repetition.
In this episode of Neural Compass with Mark Jacobstein, he interviews Vijay Mohan K. Namboodiri, who leads a research lab at UCSF where he explores the intersection of computational logic and the physical wiring of the brain.
His work investigates the possibility that learning is a retrospective process where the brain looks backward in time to find causes. By using advanced optical sensors to track dopamine fluctuations with millisecond precision, he has uncovered a mechanism that allows animals to identify the specific events that led to a reward.
This research provides a new perspective for understanding how neural circuits prioritize information and how the brain implements its own algorithms for learning.
What You’ll Learn:
- How your brain uses "time inversion" to look backward and identify the specific cause of a reward.
- The biological evidence shows that spacing out your experiences makes each individual repetition more effective for memory.
- The function of dopamine is a signal for attention that marks specific moments as meaningful.
- How can optimizing the timing of learning intervals reduce the performance differences between individual learners?
Neural Compass is a podcast from Jimini Health, which builds clinically embedded AI for behavioral health.
If you found this conversation valuable, please leave a rating or review on Apple Podcasts.
About the Guest:
Vijay Mohan K Namboodiri is a professor at UCSF, where he leads the Namboodiri Lab. His research focuses on the algorithmic and implementation levels of how animals learn associations in their environment. He specifically investigates the role of dopamine as a neuromodulator using cutting-edge recording techniques to understand the computational mechanisms of the brain.
Episode Highlights
[00:13:06] Rethinking How The Brain Learns
Professor Namboodiri challenges the dominant prediction-error model with an alternative based on “retrodictions” - learning by looking backward to identify what preceded a meaningful outcome. He explains how these backward-looking memories can then be converted into predictions about what may happen in the future.
Professor Namboodiri challenges the dominant prediction-error model with an alternative based on “retrodictions” - learning by looking backward to identify what preceded a meaningful outcome. He explains how these backward-looking memories can then be converted into predictions about what may happen in the future.
[00:17:08] Dopamine As A Signal To Pay Attention
Rather than simply treating dopamine as a reward prediction error signal, Professor Namboodiri describes it as a signal that something meaningful has happened. The increase in dopamine may prompt the brain to pay attention, look backward, and learn what caused that important event.
Rather than simply treating dopamine as a reward prediction error signal, Professor Namboodiri describes it as a signal that something meaningful has happened. The increase in dopamine may prompt the brain to pay attention, look backward, and learn what caused that important event.
[00:23:36] Why More Repetition May Not Mean More Learning
Professor Namboodiri explains a surprising prediction from his team’s retrospective learning model: the number of repetitions may have no influence on overall associative learning over a fixed period. Their experiments explore whether fewer, more widely spaced experiences can produce as much learning as many closely spaced ones.
Professor Namboodiri explains a surprising prediction from his team’s retrospective learning model: the number of repetitions may have no influence on overall associative learning over a fixed period. Their experiments explore whether fewer, more widely spaced experiences can produce as much learning as many closely spaced ones.
[00:32:02] How Spacing Could Change Learning Efficiency
The research found that across a 20-fold range of spacing, mice achieved similar overall learning despite receiving dramatically different numbers of pairings. Professor Namboodiri also explains that wider spacing reduced variability between individual animals, raising intriguing questions about whether the right pacing could reduce differences in learning ability.