Subtractive and Divisive Methods for Isolating Neural Oscillations From Broadband Activity Yield Different, Even Conflicting, Power Estimates

Understanding the role of oscillatory brain activity in cognition is a key objective in cognitive neuroscience. However, quantifying the power of narrowband neural oscillations is made challenging by the presence of non-oscillatory broadband activity in the signal. Recently, there has been a proliferation of methods designed to separate oscillatory from non-oscillatory (aperiodic) signals. To perform this separation, these methods must model the relationship between oscillatory and aperiodic components. Crucially, however, such methods rarely explicitly clarify whether they assume that this relationship is additive or multiplicative, an issue compounded by the common use of log-transformed data (i.e., subtraction in log space is equivalent to division in linear space). Here, we show the importance of justifying which assumption about component relations is applied. Using simulated power spectra with additive oscillatory and aperiodic components, we illustrate that methods that assume a multiplicative model, such as the popular specparam (formerly fooof) toolbox, systematically mis-estimate oscillatory activity in ways that can obscure real oscillatory differences between conditions, or even reverse their sign, creating an illusory effect. We then demonstrate these effects in two complementary empirical datasets: (a) a cross-sectional, between-subjects analysis of age-related differences in alpha power, showing how separation approaches impact effect size and detectability; and (b) a within-subjects analysis of pre-stimulus changes in alpha power in a detection task, showing how switching between approaches can greatly reduce the evidence for an effect, leading to different conclusions. These results highlight that researchers must explicitly report and justify assumptions about aperiodic and oscillatory signal combination, as these assumptions have implications not only for analyses of oscillatory activity and resultant conclusions, but also for integration of results across reports (which may implicitly employ different assumptions), and ultimately for the development of theories regarding electrophysiological features and their putative relations to cognition.

Authors

Institutions

Publication Details

Journal
bioRxiv (Cold Spring Harbor Laboratory)
Published
2026-11-17
DOI
https://doi.org/10.64898/2026.11.14.766643
Primary Topic
Neural dynamics and brain function
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Subtractive and Divisive Methods for Isolating Neural Oscillations From Broadband Activity Yield Different, Even Conflicting, Power Estimates

Gabriele Gratton, Bradley Voytek, Máté Gyurkovics, Grace M. Clements et al.
bioRxiv (Cold Spring Harbor Laboratory)
Neural dynamics and brain function
preprint

Subtractive and Divisive Methods for Isolating Neural Oscillations From Broadband Activity Yield Different, Even Conflicting, Power Estimates

Gabriele Gratton, Bradley Voytek, Máté Gyurkovics, Grace M. Clements, Thomas Donoghue, Monica Fabiani, Diane M. Beck, Emily Cunningham, Patrycja Kałamała, Kathy A. Low
preprint en

Abstract

Understanding the role of oscillatory brain activity in cognition is a key objective in cognitive neuroscience. However, quantifying the power of narrowband neural oscillations is made challenging by the presence of non-oscillatory broadband activity in the signal. Recently, there has been a proliferation of methods designed to separate oscillatory from non-oscillatory (aperiodic) signals. To perform this separation, these methods must model the relationship between oscillatory and aperiodic components. Crucially, however, such methods rarely explicitly clarify whether they assume that this relationship is additive or multiplicative, an issue compounded by the common use of log-transformed data (i.e., subtraction in log space is equivalent to division in linear space). Here, we show the importance of justifying which assumption about component relations is applied. Using simulated power spectra with additive oscillatory and aperiodic components, we illustrate that methods that assume a multiplicative model, such as the popular specparam (formerly fooof) toolbox, systematically mis-estimate oscillatory activity in ways that can obscure real oscillatory differences between conditions, or even reverse their sign, creating an illusory effect. We then demonstrate these effects in two complementary empirical datasets: (a) a cross-sectional, between-subjects analysis of age-related differences in alpha power, showing how separation approaches impact effect size and detectability; and (b) a within-subjects analysis of pre-stimulus changes in alpha power in a detection task, showing how switching between approaches can greatly reduce the evidence for an effect, leading to different conclusions. These results highlight that researchers must explicitly report and justify assumptions about aperiodic and oscillatory signal combination, as these assumptions have implications not only for analyses of oscillatory activity and resultant conclusions, but also for integration of results across reports (which may implicitly employ different assumptions), and ultimately for the development of theories regarding electrophysiological features and their putative relations to cognition.

bioRxiv (Cold Spring Harbor Laboratory)
University of East Anglia (GB), Jagiellonian University (PL), United States Air Force Research Laboratory (US), University of Illinois Urbana-Champaign (US), University of Michigan (US), Wright-Patterson Air Force Base (US), University of Manchester (GB), University of California San Diego (US), Wright Laboratory (US)
Neural dynamics and brain function
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon