EEG ‘Fastball’ shows promise for spotting early memory loss in Alzheimer’s

By Dr. Liji Thomas, MDReviewed by Lauren HardakerSep 8 2025

A simple EEG scan can reveal hidden memory problems in just three minutes, giving doctors a powerful new tool to detect dementia earlier and more accurately.

Study: A passive and objective measure of recognition memory in mild cognitive impairment using Fastball memory assessment. Image credit: Svitlana Hulko/Shutterstock.com

Dementia is a contributor to disability and a leading cause of mortality. Alzheimer’s disease (AD) is a major cause, accounting for 32 million cases globally. A recent paper in Brain Communications presents Fastball, a new electroencephalography (EEG) method that aims to measure recognition memory and diagnose mild cognitive impairment, an interim stage between normal cognition and dementia that is often an early manifestation of AD.

Introduction

Up to 75% of dementia is caused by AD, indicating the urgent need for better and earlier diagnostic tools. These would facilitate earlier treatment and enable patients to make proper arrangements for future care once the disease progresses. Amnesia is a risk marker for patients with mild cognitive impairment (MCI) to progress to dementia.

Amnestic MCI patients mainly suffer memory loss, while non-amnestic MCI patients have standard memory despite issues with attention or visuospatial processing. The former progress to dementia at 10%-17% a year, with 38% showing signs of dementia at 2.5 years and up to 60% after four years.

Neuropsychological testing provides subjective and objective measures of general cognitive function, attention, visual recognition, and visual acuity. Biomarkers like amyloid and tau protein deposition are used for biological AD testing, but they are expensive, invasive, and restricted in availability. Unfortunately, amyloid deposition is not inevitably a sign of cognitive decline, and vice versa.

Blood biomarkers promise earlier detection of these abnormal proteins and could help predict changes in brain structure. Researchers seek reliable methods to judge cognitive function based on these brain markers.

Cognitive testing

Cognitive impairment symptoms appear only 10-20 years after amyloid-tau deposition begins to occur. Cognitive testing helps assign clinical stages, but various demographic and cultural characteristics, including education and culture, can bias the results. Language impairments make these tests even less reliable and more complicated to apply.

A sensitive and unbiased measure of cognitive function is needed, one that can be combined with amyloid-tau biomarkers to improve both early diagnosis and accurate staging of the disease. EEG directly measures brain activity during tasks that involve thought. It records alterations in AD patients compared with controls, most often in the P300 response. This brainwave signal typically appears about 300 milliseconds after a stimulus and reflects attention and memory updating processes.

Still, it is not widely adopted to diagnose or screen individual patients because of the long duration of the test and variability between subjects.

Fastball

Fastball (fast periodic visual stimulation, FPVS) is an EEG method that can potentially be a sensitive, easily administered, inexpensive test for AD. The authors have previously shown its ability to identify AD patients compared to controls based on cognitive dysfunction. Fastball benefits include the short recording time, a high signal-to-noise ratio, and a wide range of data capture. It measures recognition memory passively, without needing behavioral memory responses or task comprehension.

Healthy older adults typically show impaired recollection of details but preserve familiarity, meaning they often cannot recall exact information yet still recognize a stimulus as previously encountered. Both AD and MCI patients have impaired recollection, while familiarity trends are more variable.

Subjective behavioral tests of familiarity are complex because language and related areas of cognition heavily influence them. However, amnestic MCI is more consistently associated with familiarity deficits. This area is selectively impacted in amnestic MCI patients who are developing dementia. Hence, it is vital to validate Fastball, which promises to assess recognition memory reliably.

About the study

The researchers tested amnestic dysfunction (failure of memory recall) in MCI patients. The study included 54 healthy older people and 53 MCI patients.

All participants completed a three-minute task consisting of watching images flashing by as Fastball monitored them. The EEG recorded how well they could automatically differentiate previously seen images from new, unfamiliar ones.  

For comparison, various neuropsychological tests were performed to examine whether Fastball responses were selective for memory (versus attention). All tests, including Fastball, were repeated after a year to validate Fastball's reproducibility and capture its sensitivity to reduced cognitive function after one year.

Study findings

Amnestic MCI patients had reduced Fastball responses compared with non-amnestic patients and healthy older adults. Analysis of their responses showed that the alterations in Fastball responses selectively identified patients at risk of losing recognition memory.

In contrast, behavioral accuracy predicted both memory and attention performance, showing it was less selective than the EEG-based Fastball measure.  

While amnestic MCI patients had less accurate recognition responses than healthy older people, the latter were comparable to non-amnestic MCI participants. For reaction times, amnestic MCI patients were slower than both groups. In comparison, non-amnestic patients were slower than controls only in some tasks (e.g., fixation cross detection) but not in oddball recognition.

The results showed moderate-to-good test-retest reliability, with Fastball outperforming the ACE-III memory subscale in healthy controls. Fastball also picked up a trend towards lower Fastball recordings at baseline in the MCI patients who developed dementia over the follow-up year (n=6). However, the authors caution that numbers were small and too low for meaningful statistical comparison, so this cannot yet be taken as predictive, pending further follow-up.

This pilot study thus established Fastball as a method to test recognition memory in patients with cognitive impairment.

Conclusions

“Fastball is a viable functional biomarker that can help to advance cognitive assessment in MCI.” It is convenient, non-invasive, easy to use, inexpensive, and scalable, not requiring the patient’s cooperation. The responses are reliable, with little variability across sessions.    

The authors say, “We do not propose that Fastball could, or even should, replace standard pen and paper cognitive testing. Rather we propose that an implicit neural measure has many potential practical advantages and with sufficient technological development could be delivered and interpreted as easily as an electrocardiogram, which is routinely delivered economically, at scale, with minimal training required.”      

Further follow-up is planned to establish the prognostic value for AD when Fastball is used as a baseline measure.

Download your PDF copy now!