Mice gradually learn an exit’s location during training but later fail to retain it: something may go wrong after acquisition. A Molecular Neurodegeneration study published on April 14, 2023 accumulates human tau in the mouse medial septum. Spatial memory declines despite comparable training performance. The investigation points to cholinergic cells connected to hippocampal CA1 and to rhythmic coordination after learning. Helping the circuit depends on both when its activity is increased and the rhythm used.

01 / Changes emerge over time, separating learning from retention

Changes emerge over time, separating learning from retention

The medial septum, part of the basal forebrain, supplies important cholinergic input to the hippocampus. Abnormal tau accumulation and cholinergic changes both occur in Alzheimer’s disease, motivating an examination of their relationship within this region over time.

With broader regional tau expression, ChAT-positive counts, hippocampal cholinergic fibers and task performance show no significant changes at three months. At six months, marker-positive counts and fiber intensity decline. Training performance remains comparable, but later testing shows slower target localization, fewer correct pokes and less target occupancy, without significantly different distance traveled.

Original figure 1Changes emerge over time, separating learning from retention

02 / Focusing on cholinergic cells

Focusing on cholinergic cells

When tau accumulation is concentrated in ChAT-associated cholinergic cells, reduced marker-positive counts and weaker CA1 cholinergic fibers appear by three months. Changes emerge earlier than at the corresponding regional-expression time point, suggesting that the targeted cellular population matters.

Tau accumulation in this population again leaves training performance without a significant group difference, while spatial memory tests one day and one week later worsen. Initial acquisition and subsequent retention become dissociated in these mice.

Original figure 2Focusing on cholinergic cells

03 / Cholinergic cells have different electrical profiles

Cholinergic cells have different electrical profiles

Recorded cholinergic cells fall into symmetric- and asymmetric-firing profiles, with the latter more common. Tau accumulation reduces input-driven firing in asymmetric cells, without significantly changing excitability in symmetric cells. A population often considered as one group contains different vulnerabilities.

The difference chiefly concerns how many spikes cells generate in response to input. Action-potential amplitude, half-width, capacitance and resting potential do not change significantly, indicating selective electrical alterations rather than deterioration of every measured property.

Original figure 3Cholinergic cells have different electrical profiles

04 / The CA1-connected branch is more vulnerable

The CA1-connected branch is more vulnerable

The septum connects with multiple hippocampal regions. Projection tracing reveals that the asymmetric-cell excitability change appears in the CA1-connected population, whereas corresponding CA3-connected cells show no significant change. Symmetric-firing cells in either tested pathway likewise lack this alteration.

Combining molecular identity, firing profile and projection destination reveals the vulnerable branch. The investigation moves from a brain region toward a connected population with a particular activity profile.

Original figure 5The CA1-connected branch is more vulnerable

05 / Rhythms remain, but coordination weakens

Rhythms remain, but coordination weakens

Two regions can retain activity while coordinating less effectively. Simultaneous field recordings reveal reduced power in parts of the septal theta range, without significant corresponding CA1 power differences. Theta coherence between the regions falls before, during and after the task.

Coherence describes rhythmic relationships between signals rather than acetylcholine release. Local CA1 rhythms remain while their relationship with septal activity weakens, directing attention toward consolidation after learning and the possible importance of coordinated communication.

Original figure 6Rhythms remain, but coordination weakens

06 / Helping retention depends on stage and rhythm

Helping retention depends on stage and rhythm

The initially tested activation condition during retrieval does not improve performance. Moving activation to immediately after new learning, under a theta-rhythmic condition, improves later target localization, correct pokes and target occupancy. Two faster rhythms do not significantly improve performance. Delaying the same theta condition until three hours after learning also yields no significant benefit.

These comparisons link the benefit to early consolidation. They demonstrate sensitivity to the particular tested timings and rhythms rather than defining a universal memory clock for every task.

Original figure 4Helping retention depends on stage and rhythm

07 / Localizing the benefit to hippocampal input

Localizing the benefit to hippocampal input

Increasing septal cholinergic input within CA1 during early consolidation also improves later spatial memory. Local blockade of the tested acetylcholine receptors removes this benefit, while distance traveled remains without a significant difference. Cholinergic signaling therefore has experimental support as part of the circuit effect.

Local human-tau accumulation in mice does not encompass the full human disease, and reduced labeling does not resolve every cell’s fate. The study locates memory vulnerability at the intersection of firing profile, projection destination and post-learning stage—three levels that may all matter when supporting an affected network.

Original figure 7Localizing the benefit to hippocampal input