Amyloid pathology compresses dynamic range and degrades spatial coding in an Alzheimer's mouse model

Technical Summary: Amyloid Pathology Compresses Dynamic Range and Degrades Spatial Coding in an Alzheimer's Mouse Model

Problem Statement
Alzheimer's disease (AD) is characterized by the disruption of neural circuits essential for memory and cognition, particularly spatial and episodic memory. While molecular and neuropathological correlates of AD are well-characterized, the specific mechanisms by which these pathologies alter the functional capacity of neural circuits to process, store, and recall information remain poorly understood. Previous studies have reported conflicting results regarding neuronal activity in AD models: some report hyperactivity in anesthetized mice, while others report hypoactivity or no change in behaving mice. Furthermore, it is unclear how aberrant excitability translates into deficits in spatial coding and whether these deficits arise locally near amyloid plaques or spread globally across the network. This study addresses the need to reconcile these discrepancies and elucidate the circuit-level basis of cognitive decline in AD, specifically focusing on the hippocampal CA1 region, which is critical for spatial memory and one of the earliest sites of amyloid pathology.

Methodology
The authors utilized the 5xFAD mouse model, which carries five familial AD mutations and exhibits early intracellular Aβ accumulation (starting at 1.5 months) and widespread plaque deposition (by 6 months). The study compared two age groups: young mice (2.1–3.0 months) and old mice (6.4–10.1 months), alongside wild-type (WT) controls.

Key experimental and analytical approaches included:

• In Vivo Two-Photon Calcium Imaging: Head-fixed mice navigated a "floating" circular track equipped with multi-modal visual and tactile cues to minimize the spatial selectivity impairments often seen in virtual reality setups. Neurons in the hippocampal CA1 were labeled with GCaMP6s or jGCaMP7s.
• Amyloid Plaque Mapping: Amyloid plaques were visualized in three dimensions using Methoxy-X04 staining approximately one week after neuronal imaging, allowing for the correlation of neuronal activity with the 3D proximity to plaques.
• Behavioral Paradigm: Mice were trained to run on the track for 6–7 days. Imaging sessions included both familiar and novel track environments to assess recall and encoding, respectively.
• Data Analysis: Neural activity was extracted using non-negative matrix factorization and deconvolved using the OASIS algorithm. The study analyzed baseline firing, locomotion-driven firing, speed tuning curves, dynamic range, network synchrony (Pearson cross-correlation), spatial information rates, spatial coherence, place field stability, and response variability. Statistical significance was assessed using hierarchical bootstrap analysis to account for the nested structure of the data (neurons within sessions within mice).

Key Contributions and Results

1. Reduced Neuronal Dynamic Range:
   The study provides the first direct experimental evidence of a reduced dynamic range in an AD mouse model. In old 5xFAD mice, baseline activity during rest was elevated compared to WT controls, while locomotion-driven firing was significantly suppressed. This combination resulted in a compressed dynamic range, limiting the neuron's ability to modulate responses to behavioral states. This effect was most pronounced in old 5xFAD mice and was observed across all plaque distances, though strongest near plaques (<20 µm).

2. Spatial and Speed Coding Deficits:

   • Speed Coding: Old 5xFAD mice exhibited a lower fraction of speed-modulated cells and a flatter population speed tuning curve compared to WT mice.
   • Spatial Coding: Spatial information rates were significantly reduced in old 5xFAD mice. This was accompanied by lower spatial coherence, reduced place-tuning dynamic range, and increased trial-to-trial response variability (indicated by a steeper variability exponent).
   • Place Fields: Old 5xFAD mice had a smaller fraction of place cells and smaller place field sizes compared to old WT mice. These deficits were most severe in neurons located near amyloid plaques.

3. Altered Network Synchrony:
   Network dynamics were disrupted in a state-dependent manner. Young 5xFAD mice showed increased synchrony (hypersynchrony) during rest compared to WT, which disappeared during locomotion. Conversely, old 5xFAD mice exhibited significant hyposynchrony (reduced synchrony) during locomotion.

4. Impaired Learning and Recall:
   Place fields in old 5xFAD mice emerged more slowly in both familiar and novel environments. In familiar tracks, this indicates impaired recall of established spatial maps. In novel tracks, it suggests a deficit in the encoding of new spatial representations. Young 5xFAD mice showed only modest, non-significant delays in tuning acquisition.

5. Progression of Pathology:
   The study demonstrates a progression of dysfunction. In young 5xFAD mice, circuit alterations (elevated baseline activity, reduced dynamic range) were largely confined to neurons in the immediate vicinity of amyloid plaques. In old 5xFAD mice, these abnormalities became more widespread across the network, although the strongest effects remained near plaques.

Significance and Claims
The authors claim that their study offers new insights into the circuit basis of cognitive decline in AD by linking local plaque pathology to impaired flexible modulation of CA1 activity. The central finding is that amyloid pathology does not merely cause global hyperactivity or hypoactivity but fundamentally compresses the dynamic range of neuronal activity. This compression limits the capacity of neurons to discriminate stimulus intensities and modulate responses based on behavioral states (e.g., rest vs. locomotion), thereby degrading information coding.

The paper posits that these deficits in dynamic range and spatial coding, which progress from local plaque-adjacent dysfunction to widespread network disruption, provide a mechanistic explanation for the memory and learning impairments observed in AD. The study concludes that understanding these specific circuit-level alterations is critical for developing therapeutic approaches, noting that the observed deficits in both recall and encoding suggest that amyloid pathology affects multiple stages of hippocampal spatial processing. The authors maintain a modest tone, framing their work as providing "direct experimental evidence" for reduced dynamic range and "new insights" into circuit dysfunction, rather than claiming to have solved the broader therapeutic challenges of AD.