Nature paper maps qEEG signatures separating Alzheimer's from FTD

A Nature paper reports quantitative EEG signatures that distinguish Alzheimer's disease from frontotemporal dementia, using spectral power and functional connectivity analyses across the two patient cohorts. The work offers clinical neurophysiology laboratories an electrophysiological handle on a differential that, in its early years, often defeats bedside assessment.

What is the diagnostic problem?

Frontotemporal dementia and Alzheimer's disease present with overlapping memory, language, and behavioral symptoms, particularly in the first two to three years of illness. Frontotemporal dementia accounts for an estimated 10 to 20 percent of all dementia cases, and is the second most common cause of early-onset dementia after Alzheimer's disease.

Structural imaging helps once atrophy is pronounced. In younger patients and at early disease stages, however, the MRI can look deceptively normal. The field has searched for years for a non-invasive, scalable biomarker that complements PET and lumbar CSF studies. EEG is cheap, repeatable, and already embedded in most memory clinics.

What did the authors set out to compare?

The investigators compared quantitative scalp EEG features between patients with Alzheimer's disease and those with frontotemporal dementia. They extracted two families of measures: spectral power across canonical frequency bands, and functional connectivity estimated from phase or amplitude relationships between electrode pairs.

The methodology treats the EEG as a multivariate pattern. It moves beyond the qualitative visual read that clinical EEG reports have historically delivered.

Why combine power and connectivity?

Power alterations in dementia are well documented in the literature. Slowing of the dominant posterior rhythm, increases in relative theta and delta power, and reductions in alpha and beta amplitude are consistent findings in Alzheimer's disease cohorts.

Connectivity analyses ask a different question: how do distant regions co-vary in time? Functional connectivity in the alpha band tends to weaken in Alzheimer's disease, while frontotemporal dementia shows partly distinct patterns that reflect its frontal and anterior temporal pathology.

Both classes of measure run on routine scalp recordings with standard montages. Neither requires intracranial electrodes nor experimental hardware.

What does the title actually claim?

The title frames the paper as a comparison of signatures, not as the introduction of a diagnostic test. Readers should treat the contribution as evidence that scalp electrophysiology carries differential information between these two dementias.

No FDA-cleared qEEG biomarker for the dementia differential currently exists. The study does not propose one. Any clinical translation would require prospective validation, ideally against PET-confirmed cohorts, plus standardization across vendors and montages.

Where does this sit in the literature?

Quantitative EEG in dementia is not new. Studies from the 1990s onward have catalogued slowing and loss of complexity in Alzheimer's disease. More recent work has used machine learning to classify patients versus controls.

The current paper extends that line of work to the harder problem of differential diagnosis. Signatures must be specific enough to survive the phenotypic noise of two diseases with shared symptoms.

It also illustrates a methodological shift. Earlier studies tended to focus on one metric at a time. Multivariate pattern analysis, by contrast, treats power and connectivity as joint features and asks which combination separates the cohorts.

What does this change for the EEG laboratory?

For a clinical neurophysiology service, the immediate takeaway is methodological. Pattern-level qEEG features can stratify patients who present with a dementia syndrome.

This matters for research workflows, for case selection in therapeutic trials, and for tertiary memory clinics running scalp EEG alongside MRI and PET. Whether such signatures translate into a billing code, a triage rule, or a reportable finding in the reading room is a question for the next round of validation, not for next quarter's dictation template.

Will this change routine practice?

Not tomorrow. The clinical EEG report stays the same for now.

The next generation of quantitative reports EEG laboratories offer to memory clinics will look different.

Probably, if the cohorts replicate in independent hands.