Start where the evidence is strongest.
The memory layer stands on decades of published memory research. Here is what we build on, and what we still need to learn.
Read the researchWhat science already shows
01
Memory loss can be lost access
In mouse models of early Alzheimer's disease, direct activation of the cells holding a memory brought it back. The problem was retrieval, not storage. Roy et al., Nature, 2016.
02
Memory can be written
In 2013, Ramirez, Liu and Tonegawa at MIT created a false memory in mice by activating the hippocampal cells that hold a memory (Science, 2013). We take one lesson from this: memory can be changed from outside, so it must be protected.
03
Human memory can be improved with stimulation
A hippocampal prosthesis that used patients' own neural codes improved memory by about 35%. Hampson et al., Journal of Neural Engineering, 2018. Closed loop stimulation of the lateral temporal cortex increased recall by about 15%. Ezzyat et al., Nature Communications, 2018.
04
Memory drifts
The neurons that represent a memory change over days and weeks, even when behavior stays the same. Ziv et al. (Nature Neuroscience, 2013) tracked thousands of hippocampal cells in mice and found that only about 15 to 25% overlapped between any two days. This is called representational drift. A memory must be modeled continuously, not saved once.
Research programs
01
Neural foundation models for memory
Models trained on human intracranial and implant recordings that decode memory state and location.
02
Human memory atlas
A shared map of where and how memories appear in human neural data.
03
Memory verification benchmark
A before and after test for brain procedures, designed with clinicians.
04
Closed loop restoration signals
Decoders that tell stimulation devices when and where to act.
05
Continuity through brain renewal
Methods to carry memory through gradual brain tissue replacement, in step with programs such as ARPA-H's Functional Repair of Neocortical Tissue (FRONT) program, launched in July 2025.
Open questions
How stable is a decoded memory across months?
How much recording is enough to verify memory?
Which signals predict that a memory can still be reached?
We will publish what we learn, including what does not work.
[Publications]
Publications
Our first results will appear here. Until then, we point to the work we build on.
Oct 2026
Why we moved from genetics to memory
We stopped asking which pill makes people live longer and started asking what has to survive for a longer life to still be theirs.
Chiron Discovery Team
Company
2020
Memory engrams: recalling the past and imagining the future
A review of how the physical trace of a memory is found, switched on and linked across brain regions.
Josselyn and Tonegawa, Science
Research
2018
Developing a hippocampal neural prosthetic to facilitate human memory encoding and recall
Replaying a patient's own hippocampal patterns improved memory by about 35% in a small group of epilepsy patients.
Hampson et al., Journal of Neural Engineering
Research
2018
Closed loop stimulation of temporal cortex rescues functional networks and improves memory
Stimulation timed to predicted memory lapses increased recall by about 15%.
Ezzyat et al., Nature Communications
Research
2016
Memory retrieval by activating engram cells in mouse models of early Alzheimer's disease
Memories that mice could not recall came back when the cells holding them were activated directly. The study behind the idea of lost access.
Roy et al., Nature
Research
2013
Creating a false memory in the hippocampus
Reactivating the cells of one memory during a new event wrote a memory of something that never happened. The reason memory must be protected.
Ramirez et al., Science
Research
2013
Long term dynamics of CA1 hippocampal place codes
The cells that code a place change from day to day while the memory holds. The first clear picture of representational drift.
Ziv et al., Nature Neuroscience
Research
Press enquiries: press@chirondiscovery.com
Research with us
If you run a memory lab, an epilepsy monitoring unit or a brain repair program, we want to talk.
Propose a collaboration