The effect of aperiodic components in distinguishing Alzheimer's disease from frontotemporal dementia

Abstract

Background: Distinguishing Alzheimer's disease (AD) from frontotemporal dementia (FTD) poses a clinical challenge, and to address this, inexpensive and accessible techniques, such as electroencephalography (EEG), are increasingly being employed. However, current studies on electrophysiology have not adequately separated aperiodic activity from spectral analysis and have instead evaluated the combination of periodic oscillatory components and aperiodic components. As the understanding of aperiodic activity has evolved, its potential physiological significance in these disorders has become an area of growing interest. The objective of our study is to determine the differences in aperiodic activity between these two dementia-related diseases and to evaluate their effect in distinguishing between them. Methods: A total of 88 participants, including 36 patients with AD, 23 patients with FTD, and 29 healthy controls (CN), were enrolled for cognitive assessment and scalp EEG acquisition. The spectrum was decomposed using a method of parameterizing neuronal power spectra, comparing group differences in different components, and a support vector machine was used to determine the effect of aperiodic parameters in the differential diagnosis. Results: The AD and FTD groups showed varying degrees of EEG rhythm slowing compared to CN group. Theta periodic power and aperiodic parameters were higher in the AD group compared to the FTD group at the channel level. The inclusion of the aperiodic parameters led to better performance for differential diagnosis. Conclusions: Our study indicates that aperiodic activity shows differences in patients with AD and FTD, and has the potential to distinguish between these disorders. These results underscore the physiological significance of aperiodic components in dementia-related diseases.

A holo-spectral EEG analysis provides an early detection of cognitive decline and predicts the progression to Alzheimer’s disease

“…If the results in our data and Wang et al (2023) are reliable, then the alpha band is likely a fulcrum balancing the lower-frequency (LFO; delta and theta) power and higher-frequency (HFO; beta and gamma) power. The result shows increased aperiodic exponent in AD patients, which means the slope is steeper, leading to increasing LFO power and decreasing HFO power ( Figure 3C ; Wang et al, 2023 ). In terms of excitation/inhibition (E/I) imbalance, it depicts a hypoexcitability state ( Gao et al, 2017 ), which is not the case in the AD animal model of Palop and Mucke (2016) , which is also contrary to a recent study showing that AD patients have a lower aperiodic gamma frequency exponent and decreased DFA exponent, which suggests a flatter slope (i.e., hyperexcitability state) and an E/I imbalanced state, respectively ( van Nifterick et al, 2023 ).…”

“…The changing trend in the center frequency is consistent in the aging and AD process. In contrast, the trend in aperiodic parameters (i.e., offset and exponent) is the opposite in both conditions, decreases across aging ( Donoghue et al, 2020 ; Merkin et al, 2023 ) and increases in AD ( Figure 3E ; Wang et al, 2023 ). If the results in our data and Wang et al (2023) are reliable, then the alpha band is likely a fulcrum balancing the lower-frequency (LFO; delta and theta) power and higher-frequency (HFO; beta and gamma) power.…”

“…In contrast, the trend in aperiodic parameters (i.e., offset and exponent) is the opposite in both conditions, decreases across aging ( Donoghue et al, 2020 ; Merkin et al, 2023 ) and increases in AD ( Figure 3E ; Wang et al, 2023 ). If the results in our data and Wang et al (2023) are reliable, then the alpha band is likely a fulcrum balancing the lower-frequency (LFO; delta and theta) power and higher-frequency (HFO; beta and gamma) power. The result shows increased aperiodic exponent in AD patients, which means the slope is steeper, leading to increasing LFO power and decreasing HFO power ( Figure 3C ; Wang et al, 2023 ).…”

“…Similarly, alpha AM power spectra fitted in the FOOOF algorithm yield periodic components with two peak frequencies at 3 and 8 Hz and similar aperiodic exponents of all groups ( Supplementary Figure 15 , middle and right column). As Wang et al (2023) revealed in the adjusted power spectrum, the alpha periodic component revealed decreased power in AD compared to CN, with an increased aperiodic offset and exponent ( Wang et al, 2023 ). The changing trend in the center frequency is consistent in the aging and AD process.…”

Oscillatory characteristics of resting-state magnetoencephalography reflect pathological and symptomatic conditions of cognitive impairment

“…Similar to classical studies, this study assumed that changes in the neural power spectra reflected the changes in oscillatory activities associated with cognitive impairment. However, recent studies have demonstrated the contribution of aperiodic component as a predictor of cognitive impairment, which found that the aperiodic components differed between AD vs. FTD ( Wang et al, 2023 ) and DLB/PD vs. MCI/control ( Rosenblum et al, 2023 ), while other studies showed no differences in the aperiodic components among healthy controls, patients with MCI, and patients with AD ( Azami et al, 2023 ; Kopčanová et al, 2023 ). In examining the correlations between RPs and NPAs, this study demonstrated that the correlational directions were flipped between low- (from RPd to RPa1) and high-frequency bands (from RPa3 to RPhg), with a border transition at RPa2 ( Figures 1 , 4 , 5 ).…”