Band ratio measures, computed as the ratio of power between two frequency bands, are a common analysis measure in neuro-electrophysiological recordings. Band ratio measures are typically interpreted as reflecting quantitative measures of periodic, or oscillatory, activity. This assumes that the measure reflects relative powers of distinct periodic components that are well captured by predefined frequency ranges. However, electrophysiological signals contain periodic components and a 1/f-like aperiodic component, the latter of which contributes power across all frequencies. Here, we investigate whether band ratio measures truly reflect oscillatory power differences, and/or to what extent ratios may instead reflect other periodic changes-such as in center frequency or bandwidth-and/or aperiodic activity. In simulation, we investigate how band ratio measures relate to changes in multiple spectral features, and show how multiple periodic and aperiodic features influence band ratio measures. We validate these findings in human electroencephalography (EEG) data, comparing band ratio measures to parameterizations of power spectral features, and find that multiple disparate features influence ratio measures. For example, the commonly applied theta / beta ratio is most reflective of differences in aperiodic activity, and not oscillatory theta or beta power. Collectively, we show that periodic and aperiodic features can create the same observed changes in band ratio measures, and that this is inconsistent with their typical interpretations as measures of periodic power. We conclude that band ratio measures are a non-specific measure, conflating multiple possible underlying spectral changes, and recommend explicit parameterization of neural power spectra as a more specific approach. Band Ratios 3 Significance Statement Neural oscillations are a ubiquitous feature of investigation in electrophysiological recordings. Frequency band ratio measures are a common approach to investigate neural oscillations, applied across cognitive and clinical neuroscience, and in recording modalities such as in electroencephalography and local field potentials. In this work we systematically investigate the methodological properties of band ratio measures. We show that band ratio measures are not specific to measuring oscillatory power, as they are intended and interpreted to do. Rather, they often reflect other features of the data, such as aperiodic, or 1/f-like, activity. These findings are significant for interpreting prior empirical and clinical research, guiding future work, and another motivation that aperiodic neural activity should be a key consideration when studying electrophysiological data.

Acknowledgements 32 33We would like to thank members of the Voytek Lab for insightful comments and suggestions 34 throughout this project. We would also like to express gratitude to the many people involved in 35 generating the open-access datasets and developing the open-source tools that made this project 36 possible. Abstract 39 40A common analysis measure for neuro-electrophysiological recordings is to compute the 41 power ratio between two frequency bands. Applications of band ratio measures include 42 investigations of cognitive processes as well as biomarkers for conditions such as attention-deficit 43 hyperactivity disorder. Band ratio measures are typically interpreted as reflecting quantitative 44 measures of periodic, or oscillatory, activity, which implicitly assumes that a ratio is measuring the 45 relative powers of two distinct periodic components that are well captured by predefined frequency 46 ranges. However, electrophysiological signals contain periodic components and a 1/f-like aperiodic 47 component, which contributes power across all frequencies. In this work, we investigate whether 48 band ratio measures reflect power differences between two oscillations, as intended. We examine 49 to what extent ratios may instead reflect other periodic changes-such as in center frequency or 50 bandwidth-and/or aperiodic activity. We test this first in simulation, exploring how band ratio 51 measures relate to changes in multiple spectral features. In simulation, we show how multiple 52 periodic and aperiodic features affect band ratio measures. We then validate these findings in a 53 large electroencephalography (EEG) dataset, comparing band ratio measures to parameterizations 54 of power spectral features. In EEG, we find that multiple disparate features influence ratio measures. 55For example, the commonly applied theta / beta ratio is most reflective of differences in aperiodic 56 activity, and not oscillatory theta or beta power. Collectively, we show how periodic and aperiodic 57 features can drive the same observed changes in band ratio measures. Our results demonstrate how 58 ratio measures reflect different features in different contexts, inconsistent with their typical 59 interpretations. We conclude that band ratio measures are non-specific, conflating multiple possible 60 underlying spectral changes. Explicit parameterization of neural power spectra is better able to 61 provide measurement specificity, elucidating which components of the data change in what ways, 62 allowing for more appropriate physiological interpretations. 63 64 Keywords 65 66 neural oscillations, frequency band ratios, spectral power ratios, theta / beta ratio, theta / alpha 67 ratio, alpha / beta ratio, electroencephalography, 1/f activity, aperiodic neural activity 68 69 Abbreviations 70 71 EEG: electroencephalography; MEG: magnetoencephalography; ECoG: electrocorticography; LFP: 72 local field potential; TBR: theta / beta ratio; TAR: theta / alpha ratio; ABR: alpha / beta ratio; CF: 73 center frequency; PW: power; BW: b...