A 34-year-old patient sat across from me last year, three months after a stroke that damaged a sliver of her hippocampus. She could still make decisions just fine when the options were in front of her - restaurant menu, no problem. But ask her to plan a week of meals, and she'd stare at you like you'd asked her to solve differential equations in Mandarin. Her planning hadn't vanished. It had changed shape. She could still reason about what was right in front of her, but the quiet, behind-the-scenes rehearsal that the rest of us do without noticing - the mental simulation that builds our sense of what's possible before we even need to choose - that part had gone dark.

Her case kept nagging at me, because the textbooks say planning is basically a chess engine in your skull: simulate moves, evaluate outcomes, pick the best one. But that's not what her brain had lost. And a new review from Marcelo Mattar and Nathaniel Daw argues that neuroscience has been thinking about planning wrong all along (Mattar & Daw, 2026).

Your Brain's Night Shift Is Running the Whole Show

Here's the traditional story: your hippocampus "replays" past experiences - neurons fire in fast-forward sequences while you rest or sleep - and this replay acts like a mental GPS, plotting routes before you commit to a direction. Neat. Tidy. And, according to the authors, incomplete.

Yes, hippocampal replay happens when you're deciding which way to turn in a maze. But it also fires up at seemingly random moments - while you're napping, while you're spacing out, while you're doing absolutely nothing that requires a decision. A recent review found surprisingly thin evidence that awake replay directly steers real-time choices in rodents (van der Meer & Bendor, 2025). Instead, replay seems to be teaching downstream brain circuits, training them like a coach running drills long before game day. It's less "which turn do I take now?" and more "let me quietly reorganize everything I know so future-me makes better calls."

Mattar and Daw's earlier work showed that the brain prioritizes which memories to replay based on their expected usefulness - not just recency or emotional weight, but actual computational value for improving future decisions (Mattar & Daw, 2018, PMCID: PMC6203620). Your sleeping brain is basically running a cost-benefit analysis on its own memories. Respect.

The Grid Cells That Skip the Queue

If replay is the brain's rehearsal strategy, grid cells are its cheat code. These neurons in the entorhinal cortex fire in eerily regular hexagonal patterns as you move through space - they won the Nobel Prize in 2014, which is the neuroscience equivalent of going platinum. But here's the twist: grid cells don't just map physical space. They encode abstract relationships and temporal structure too.

Mattar and Daw argue that these temporally abstract representations let the brain leap across states without laboriously simulating each step. Think of it as the difference between driving turn by turn versus glancing at a map and knowing the general direction. Grid-like codes effectively compress the problem, enabling a kind of planning that doesn't require the sequential, one-step-at-a-time search that AI algorithms traditionally rely on. Your brain found a shortcut, and it didn't even need to Google it.

When Your Prefrontal Cortex Learns How to Learn

The third piece of this puzzle might be the wildest. Metalearning - literally learning how to learn - appears to shape how the prefrontal cortex handles planning across different situations. Recent neural network models show that a system can learn general planning strategies across many tasks, then rapidly deploy the right approach in new contexts (Jensen et al., 2024, PMCID: PMC11239510). This mirrors how reinforcement learning principles in the medial prefrontal cortex help the brain build flexible, abstract schemas - simplified internal models that compress experience into reusable knowledge (Bein & Niv, 2025).

In other words, your prefrontal cortex isn't running one fixed planning algorithm. It's running the algorithm that it learned works best for situations like this one. It's adaptive all the way down.

So What Is Planning, Then?

Mattar and Daw's punchline is elegant: planning isn't one thing. It's a family of processes where mental simulation serves learning - building better internal models, training faster circuits, compressing experience into useful abstractions. The classic "simulate-then-choose" version of planning? That's just one member of the family. And maybe not even the one who shows up most often.

For my patient, this reframing actually matters. Her deficit wasn't a broken chess engine. It was a disruption to the quiet, offline machinery that builds the foundation planning rests on. The rehearsal. The reorganization. The invisible prep work.

Your brain plans while you sleep, while you daydream, while you're doing nothing at all. And now we're finally starting to understand why.

References

1. Mattar, M. G., & Daw, N. D. (2026). Planning in the Brain: It's Not What You Think It Is. Annual Review of Neuroscience. DOI: 10.1146/annurev-neuro-102124-015847

2. Mattar, M. G., & Daw, N. D. (2018). Prioritized memory access explains planning and hippocampal replay. Nature Neuroscience, 21(11), 1609-1617. DOI: 10.1038/s41593-018-0232-z. PMCID: PMC6203620

3. Jensen, K. T., Hennequin, G., & Mattar, M. G. (2024). A recurrent network model of planning explains hippocampal replay and human behavior. Nature Neuroscience, 27(7), 1340-1348. DOI: 10.1038/s41593-024-01675-7. PMCID: PMC11239510

4. van der Meer, M. A., & Bendor, D. (2025). Awake replay: off the clock but on the job. Trends in Neurosciences, 48(4), 257-267. DOI: 10.1016/j.tins.2025.02.006

5. Bein, O., & Niv, Y. (2025). Schemas, reinforcement learning and the medial prefrontal cortex. Nature Reviews Neuroscience, 26(3), 141-157. DOI: 10.1038/s41583-024-00893-z

Disclaimer: The image accompanying this article is for illustrative purposes only and does not depict actual experimental results, data, or biological mechanisms.