Generating realistic neurophysiological time series with denoising diffusion probabilistic models

Abstract

In recent years, deep generative models have had a profound impact in engineering and sciences, revolutionizing domains such as image and audio generation, as well as advancing our ability to model scientific data. In particular, Denoising Diffusion Probabilistic Models (DDPMs) have been shown to accurately model time series as complex high-dimensional probability distributions. Experimental and clinical neuroscience also stand to benefit from this progress, since accurate modeling of neurophysiological time series, such as electroencephalography (EEG), electrocorticography (ECoG), and local field potential (LFP) recordings, and their synthetic generation can enable or improve a variety of neuroscience-specific applications. Here, we present a method for modeling multi-channel and densely sampled neurophysiological recordings using DDPMs, which can be flexibly applied to different recording modalities and experimental configurations. First, we show that DDPMs can generate realistic synthetic data for a variety of datasets including different recording techniques (LFP, ECoG, EEG) and species (rat, macaque, human). DDPM-generated time series accurately capture single- and multi-channel statistics such as frequency spectra and phase-amplitude coupling, as well as fine-grained and dataset-specific features such as sharp wave-ripples. In addition, synthetic time series can be generated based on additional information like experimental conditions or brain states. We demonstrate the utility and flexibility of DDPMs in several neuroscience-specific analyses, such as brain-state classification and imputation of missing channels to improve neural decoding. In summary, DDPMs can serve as accurate generative models of neurophysiological recordings, and have a broad utility in the probabilistic generation of synthetic time series for many applications relevant to neuroscience.

Modeling conditional distributions of neural and behavioral data with masked variational autoencoders

“…More generally, data imputation can be framed as modeling conditional distributions over missing variables given observed ones. This is a relevant pre-processing step for many downstream analyses that require complete datasets in neuroscientific and clinical applications, as well as in other domains ( Talukder et al, 2022 ; Vetter et al, 2023 ). For example, if an electrode breaks during a neural recording, a masked VAE approach can salvage the dataset by computing conditionals for the failed electrode using complete data from other sessions.…”

Using cluster-based permutation tests to estimate MEG/EEG onsets: how bad is it?

“…When estimating onsets, several steps can be taken to improve and assess the precision of the results. To extend my simple simulations, one could use algorithms that generate realistic time-series of various types of effects (Barzegaran et al, 2019; Julius Vetter et al, 2023). When dealing with real data, instead of estimating onsets at the group level, it is better practice to measure onsets in each participant and to report group distributions (Bieniek et al, 2016).…”

A Pose-Informed De-Noising Diffusion Model for Adult Naturalistic EEG Signals

“…One method of surpassing this limitation is by utilising an ODE solver which may be implemented in future work. Further, our model assumes the noise to follow an isotropic Gaussian which can be improved upon by assuming the noise variance evolution to be correlated across diffusion time steps, see for instance the approach in the recent work [21] that however does not consider denoising EEG artifacts.…”