Alzheimer's disease (AD), the most common cause of dementia, is defined by amyloid plaques, neurofibrillary tangles (NFTs) of hyperphosphorylated tau, and neuronal loss. The traditional amyloid cascade hypothesis long held that amyloid-β (Aβ) accumulation is the primary instigating event, supported by early genetic evidence linking AD to mutations in Aβ-processing genes (APP, PSEN1/2).

However, this paradigm is challenged by clinical trial results. Approved anti-Aβ immunotherapies, despite significantly reducing plaque levels, confer only modest clinical benefits, slowing cognitive decline slightly. Tau-based therapies have also failed, though they have not yet successfully cleared NFTs from the perikarya of cortical pyramidal neurons.

This evidence supports a proposed shift to the cerebrocortical disconnection hypothesis. This theory posits that dementia results from the breakdown of neural networks, initiated when tau aggregates into NFTs. This disruption damages axonal transport, creating an energetic crisis at nerve terminals, which promotes the production of Aβ42 and subsequent amyloid plaque formation.

The neural breakdown is further worsened by the dysfunction and loss of myelin-maintaining oligodendrocytes, leading to demyelination. This forces the damaged axons to expend even more energy for signalling, severely exacerbating their bioenergetic deficit. This combination of faulty transport, amyloid pathology, and myelin loss ultimately destroys critical long corticocortical and corticofugal neurons, along with (cholinergic, noradrenergic and serotonergic) corticopetal neurons. We conclude that therapeutic focus must now shift toward protecting the integrity of neural circuitry directly to prevent disconnection and meaningfully slow cognitive decline.