Abstract
Emerging research indicates that gamma-band oscillations are sensitive indicators of synaptic health and are significantly altered in the early stages of Alzheimer’s Disease (AD). This study aimed to determine if a cognitive "stress test"—combining an N-back working memory task with quantitative EEG (qEEG)—could unmask physiological abnormalities in individuals who are cognitively healthy (CH) but possess pathological cerebrospinal fluid (CSF) biomarkers for AD. We analyzed low (30–50 Hz) and high (50–80 Hz) gamma power across six brain regions in 24 participants stratified into normal (CH-NAT) and pathological (CH-PAT) groups based on amyloid/tau ratios. Our findings reveal a pattern of frontocentral hyperactivity during low-load tasks and a subsequent failure to maintain gamma power during high-load challenges in the CH-PAT group. These results suggest that gamma activity is a viable non-invasive marker for identifying individuals in the prodromal stage of AD.
Introduction
The "silent" phase of Alzheimer’s Disease, occurring years before clinical symptoms manifest, represents a critical window for therapeutic intervention. However, identifying individuals in this stage typically relies on invasive CSF draws or expensive PET imaging. There is an urgent need for functional, non-invasive markers that reflect the state of synaptic health. Gamma oscillations (30–80 Hz) are particularly promising as they are generated by the activity of inhibitory GABAergic interneurons, which are known to be vulnerable in the early stages of AD pathology. This study investigates whether the combination of a cognitive challenge and qEEG can reveal early functional deficits in the "at-risk" brain.
Methods
Participants and Stratification
The study cohort consisted of 24 cognitively healthy adults aged 60 and older. Participants were categorized based on their CSF amyloid-beta and tau protein ratios:
- CH-NAT (Normal Amyloid/Tau): 10 participants with biomarker profiles within the normal range.
- CH-PAT (Pathological Amyloid/Tau): 14 participants with biomarker profiles indicative of AD risk.
Cognitive Challenge: The N-back Task
Participants performed an N-back working memory task under two conditions:
- 0-back (Low Load): A simple target detection task where participants identified a specific letter.
- 2-back (High Load): A more demanding task requiring participants to determine if the current letter matched the one presented two steps prior.
qEEG Data Acquisition and Analysis
EEG data were recorded during the tasks, and power in the low gamma (30–50 Hz) and high gamma (50–80 Hz) bands was calculated using Morlet wavelet convolution. Analysis was conducted across six predefined regions: frontal, central, parietal, left temporal, right temporal, and occipital.
Results
Behavioral Performance and Cognitive Load
While both groups showed comparable performance on the 0-back task, a significant divergence emerged during the 2-back condition. The CH-PAT group exhibited significantly lower accuracy ($p = 0.029$, Cohen’s $d = -0.87$), suggesting that the presence of AD biomarkers is associated with impaired working memory performance when cognitive demand is high.
Gamma Power: Hyperactivity and Failure
The qEEG analysis revealed a distinct neurophysiological signature in the CH-PAT group. During the low-load 0-back task, CH-PATs showed significantly higher low-gamma power in frontocentral regions compared to CH-NATs ($p = 0.014 \sim 0.032$, Cohen’s $d = 0.95 \sim 1.11$). This likely represents a compensatory but inefficient over-recruitment of neural resources.
However, during the high-load 2-back task, this compensatory mechanism failed. The CH-PAT group showed significantly lower low-gamma power in the left temporal region ($p = 0.045$, Cohen’s $d = -0.96$) and lower high-gamma power in the parietal region ($p = 0.035$, Cohen’s $d = -1.02$). Across the entire scalp, the effect sizes for gamma power in the high-load condition were medium to large and negative (Cohen’s $d = -0.51 \sim -1.02$).
Brain-Behavior Coupling
In healthy aging (CH-NAT), gamma power during the 2-back task was strongly and positively correlated with 0-back accuracy ($r = 0.68 \sim 0.78$) and negatively correlated with response times. This suggests that in normal aging, the ability to modulate gamma power is directly linked to cognitive efficiency. In contrast, this vital brain-behavior coupling was significantly disrupted in the CH-PAT group.
Discussion
The findings of this study provide preliminary evidence that gamma power is compromised in AD-biomarker positive individuals who are otherwise cognitively healthy. The observed pattern—hyperactivity during simple tasks followed by a failure to maintain power during complex tasks—aligns with the theory that amyloid and tau pathology disrupt the excitatory/inhibitory (E/I) balance of neural circuits. The loss of correlation between gamma power and behavioral performance in the CH-PAT group further underscores the functional impact of these early pathological changes.
Conclusion
This pilot study highlights the potential of combining cognitive challenges with qEEG as a sensitive, non-invasive method for identifying individuals in the prodromal stage of Alzheimer’s Disease. While the small sample size necessitates further validation in larger, more diverse cohorts, these results encourage the development of neurophysiology-based markers to improve our understanding of AD pathophysiology and enhance early detection efforts.
Summary updated