Time-Frequency Representations of Brain Oscillations: Which One Is Better?

Abstract

Brain oscillations are thought to subserve important functions by organizing the dynamical landscape of neural circuits. The expression of such oscillations in neural signals is usually evaluated using time-frequency representations (TFR), which resolve oscillatory processes in both time and frequency. While a vast number of methods exist to compute TFRs, there is often no objective criterion to decide which one is better. In feature-rich data, such as that recorded from the brain, sources of noise and unrelated processes abound and contaminate results. The impact of these distractor sources is especially problematic, such that TFRs that are more robust to contaminants are expected to provide more useful representations. In addition, the minutiae of the techniques themselves impart better or worse time and frequency resolutions, which also influence the usefulness of the TFRs. Here, we introduce a methodology to evaluate the “quality” of TFRs of neural signals by quantifying how much information they retain about the experimental condition during visual stimulation and recognition tasks, in mice and humans, respectively. We used machine learning to discriminate between various experimental conditions based on TFRs computed with different methods. We found that various methods provide more or less informative TFRs depending on the characteristics of the data. In general, however, more advanced techniques, such as the superlet transform, seem to provide better results for complex time-frequency landscapes, such as those extracted from electroencephalography signals. Finally, we introduce a method based on feature perturbation that is able to quantify how much time-frequency components contribute to the correct discrimination among experimental conditions. The methodology introduced in the present study may be extended to other analyses of neural data, enabling the discovery of data features that are modulated by the experimental manipulation.

Sharp detection of oscillation packets in rich time-frequency representations of neural signals

“…Neural computations and information transmission in the brain are accompanied by oscillations ( Wang, 2010 ) embedded in rich time-frequency landscapes ( Moca et al, 2021 ; Bârzan et al, 2022 ). Oscillations often appear as events of finite duration and finite frequency span, called oscillation bursts or packets, intermixed with sustained oscillations and transient broadband events ( Tal et al, 2020 ).…”

Interpretable many-class decoding for MEG

“…We did compare our approach with the wavelet features and found the latter to be somewhat inferior (see Inline Supplementary Figure 6). A more recent approach, termed superlets transform ( Moca et al, 2021 , Jorntell and Kesgin, 2023 ) has been shown to improve classification results by mitigating the time vs. frequency resolution problem ( Bârzan et al, 2022 ). However, a full comparison between different time–frequency features is out of the scope of this paper, as our main comparison between sliding-window and full-epoch decoding is performed at the raw data level.…”

Beta bursts question the ruling power for brain–computer interfaces

“…Each subject's data were first transformed in the time-frequency domain from 1 to 43 Hz using the superlets algorithm [98] with a frequency resolution of 0.5 Hz. We selected the superlets algorithm over other more commonly used methods as it allows us to obtain a more optimal tradeoff between temporal and spectral resolution, and because it has been shown to yield better classification results compared to other approaches [99]. Before proceeding with any further analysis we trimmed 200 to 250 ms from the beginning and end of the epoched data in order to exclude any edge effects introduced by the time-frequency transform.…”