Researchers have identified a precise role for the protein tau in stabilising long-term memories and shown how disease-associated forms of tau can block memory formation and retrieval in mice. The study, published on 17 May 2026 in Nature Communications, was led by Renée Kosonen with Kristie Stefanoska and senior author Arne Ittner.
What the team found
The researchers focused on small clusters of neurons known as engram cells, which form the physical blueprint of a memory. They report that tau is not required for the immediate formation or short-term recall of a new memory. Instead, tau’s critical role emerges later: it helps decide which experiences are kept for weeks or months.
During learning, tau undergoes a chemical modification called phosphorylation at a precise site known as T205. That modification appears to dampen surrounding neural ‘noise’, allowing the subset of engram cells active during an experience to consolidate a robust, long-lasting trace. When tau was removed in the mice, initial memory formation still occurred but the memories lacked durable access through natural pathways. Importantly, the study showed that artificially stimulating the engram cells could still bring those memories back.
“Why some memories last while others fade has long puzzled scientists.” — Arne Ittner, senior author
Disease tau and the breakdown of memory indexing
The researchers then introduced abnormal, disease-associated forms of tau into engram cells. Two distinct disruptions were observed:
- If faulty tau was present while learning was taking place, formation of new memories was halted.
- If the same disease forms of tau were added after a memory had been stored, retrieval was blocked.
Both manipulations produced chaotic, erratic patterns of brain activity, suggesting the problem was not simply erasure of stored data but a failure of the indexing system that links cues (sights, sounds) to memory traces.
Why this matters
These results shift part of the dementia conversation. Traditionally, memory loss in conditions such as Alzheimer’s disease has been framed largely as destruction of stored information — synapses dying or cells disappearing. This work suggests an additional mechanism: memories may remain encoded but be inaccessible because the molecular index that links retrieval cues to those traces is broken.
That distinction is important for therapies. If memories persist but cannot be accessed, interventions that restore the indexing function or bypass it (for example, targeted stimulation of engram cells) could recover information that appears lost. The mouse experiments showed such stimulation reinstated memories that were otherwise unreachable when tau was absent.
Caveats and next steps
Two constraints must be emphasised. First, these experiments were conducted in mice. Translating molecular and circuit findings from rodents to humans is challenging; brain organisation and disease progression differ. Second, the study identifies a specific phosphorylation site (T205) where tightly regulated, low-level modification appears beneficial, whereas pathological tau disrupts normal activity. Moving from that molecular insight to safe, effective therapies will require careful work to restore balance rather than simply suppress tau.
Future work will need to probe whether the same indexing mechanism operates in human memory circuits, whether similar phosphorylation patterns exist in people, and how disease tau interacts with other pathological processes in Alzheimer’s, such as amyloid pathology and inflammation.
| Condition | Effect on memory | Brain activity |
|---|---|---|
| Normal tau with T205 phosphorylation | Long-term stabilisation and retrieval of memories | Organised, low-noise activity around engram cells |
| Absent or disease-associated tau | Initial memory forms but long-term access impaired; formation or retrieval blocked if disease tau present | Chaotic, erratic activity that breaks indexing |
Implications for South Africa
Dementia prevalence is rising locally with an ageing population, and South African clinicians and researchers are seeking interventions that preserve cognition or restore function. The new findings add a mechanistic target — the tau-dependent indexing process — that may guide laboratory and translational studies here and abroad. However, any optimism should be measured: strengths of the work lie in its precise molecular and circuit dissection in mice; the path to human treatments remains long and uncertain.
In sum, the study provides a clearer picture of how tau contributes to memory persistence and how its pathological forms can selectively block either the formation or the retrieval of memories. It reframes parts of memory loss in dementia as a failure of indexing rather than only erasure, a distinction that could open different therapeutic strategies if confirmed in humans.
