In 2011, the President of the United States, Barack Obama, signed the National Alzheimer's Project Act, which was followed by the national plan to address and treat Alzheimer's disease (AD) in 2012. In effect, the US federal government viewed AD as a national crisis requiring immediate action. Why the national focus on AD? The major issue is that AD (1) has a high prevalence, (2) will have an increasing incidence with the aging population, (3) has an immense societal cost, and (4) lacks disease-modifying treatments.

Regarding this last point, part of the problem may lie in an overemphasis of the structural damage caused by abnormal proteins (amyloid and tau) in AD. Consistent with this, recent biological agents that target such abnormal proteins have failed to lead to cognitive and clinical improvement in human trials.1 In addition, the correlation between amyloid proteins and cognitive function remains tenuous.2 Could protein accumulation therefore be an epiphenomenon, with cognitive impairment driven by another mechanism? Indeed, a recent alternate hypothesis suggests that subclinical seizures, not proteins, might be the ultimate driving force behind the symptoms of AD.3-5

This “seizure theory” postulates that abnormal protein development in AD leads to the death of inhibitory interneurons, network hyperexcitability that lowers the seizure threshold, chronic subclinical seizures, and, finally, cognitive impairment.4 A recent study recorded extended periods of video electroencephalography (EEG) as well as magnetoencephalography in AD patients, and found that subclinical seizures (SCZs) occur at a rate of 4 times as often as in age-matched controls. In addition, AD patients with SCZs experienced a more rapid rate of cognitive impairment than those with no SCZs.6 This study presented an obvious problem: SCZs in the hippocampus are not readily recordable by surface EEG, and may serve as an important biomarker. In order to show that SCZs correlate with AD symptoms, it is necessary to record intracranial electroencephalography (iEEG) from the medial temporal lobe in AD patients. However, subjecting AD patients to a brain surgery is outside the standard of care.

Alice Lam and colleagues7 recently addressed this issue by implanting foramen ovale electrodes in AD patients. These electrodes are implanted using minimally invasive technology and avoid much of the morbidity associated with traditional brain surgery. The authors were thus able to record iEEG activity from the medial temporal lobe of 2 AD patients. Both patients showed all of the signs, symptoms, and radiological and pathological markers of early-onset probable AD. They found that each patient was having epileptic spikes at rates much higher than recorded on scalp EEG, and the authors found SCZs in 1 patient that was not recorded on scalp EEG. Critically, suppressing the SCZs with medication significantly reduced the rate of cognitive decline.

The article has a number of important implications. First, it is of immediate importance to expand these results in a larger trial with a greater number of patients. If the seizure theory is true, the number of SCZs should correlate with the rate of cognitive decline in AD, at least in a subset of patients. It may also serve as a biomarker in diagnosing different types of dementias. Second, it opens up the possibility of surgical targets for Alzheimer's disease. Indeed, if SCZ is the causative agent in the cognitive dysfunction of AD, 1 potential treatment strategy for AD might be iEEG monitoring, followed by antiepileptic drugs, and then neuromodulation or other surgery in cases refractory to medication. Finally, by lending evidence to the seizure theory of epilepsy, it represents a major contribution to AD research-future basic science work is needed to delineate the pathophysiology of network dysfunction in AD and other dementias. Whatever the ultimate impact of the seizure theory of AD, Lam and colleagues7 should he applauded for their thoughtful contribution to a pressing clinical problem.