Menu

Abstract blue neural threads. Temporary C01 thumbnail.

Currents

C01

Reward size and learning, hippocampal memories and flexible decisions, and Neuropixels 2.0 Quad Base.

Can a larger reward dramatically speed up learning? In our first episode of Currents, Rudy discusses “Reward magnitude determines reinforcement learning efficiency,” from the Dudman and Coddington labs at Janelia. We explore how increasing reward size accelerated learning across three mouse tasks, with immediate implications for how neuroscientists train animals. The findings also raise a deeper question about dopamine’s role in learning: could it help determine how quickly an animal learns? We trace the history of the reward prediction error framework and discuss what dopamine stimulation reproduced—and what it left unexplained.

How do memories help us apply the right rule to a decision? Amirreza presents “Hippocampal engrams configure prefrontal context representations to guide flexible decisions,” from the Tank and Brody labs at Princeton. In a virtual reality task, reactivating hippocampal neurons associated with one context biased mice toward following that context’s rule, even when they had been shown another. We discuss how the experiment separates context from a specific movement, how reactivation shifts prefrontal context representations, and why choosing which neurons to perturb—and when—can reveal how memories guide flexible behavior.

We also cover Neuropixels 2.0 Quad Base, which increases simultaneous recording capacity to 1,536 channels. We briefly trace the development of Neuropixels probes and discuss what this expanded capacity could offer experiments tracking activity across multiple brain regions.

References

  1. Gong, S., Martell, A., Dudman, J. T., & Coddington, L. T. (2026). Reward magnitude determines reinforcement learning efficiency (https://doi.org/10.1126/science.aeb0813). Science, 392(6800), eaeb0813.

  2. Julian, J. B., Kaminsky, J. C., Tank, D. W., & Brody, C. (2026). Hippocampal engrams configure prefrontal context representations to guide flexible decisions (https://doi.org/10.64898/2026.07.06.732916). bioRxiv [preprint].

  3. Chang, Y., et al. (2026). Neuropixels 1536 Channel Quad Base probe reveals brain-wide communication underlying flexible sensorimotor sequences (https://doi.org/10.64898/2026.07.23.740388). bioRxiv [preprint].

Background papers discussed

  1. Schultz, W., Dayan, P., & Montague, P. R. (1997). A neural substrate of prediction and reward (https://doi.org/10.1126/science.275.5306.1593). Science, 275(5306), 1593–1599.

  2. Mohebi, A., et al. (2019). Dissociable dopamine dynamics for learning and motivation (https://doi.org/10.1038/s41586-019-1235-y). Nature, 570, 65–70.