Thomas Donoghue’s research while affiliated with University of California, San Diego and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (34)


Dissociating Contributions of Theta and Alpha Oscillations from Aperiodic Neural Activity in Human Visual Working Memory
  • Preprint

December 2024

·

23 Reads

Quirine van Engen

·

Geeling Chau

·

Aaron Smith

·

[...]

·

While visual working memory (WM) is strongly associated with reductions in occipitoparietal 8-12 Hz alpha power, the role of 4-7 Hz frontal midline theta power is less clear, with both increases and decreases widely reported. Here, we test the hypothesis that this theta paradox can be explained by non-oscillatory, aperiodic neural activity dynamics. Because traditional time-frequency analyses of electroencephalopgraphy (EEG) data conflate oscillations and aperiodic activity, event-related changes in aperiodic activity can manifest as task-related changes in apparent oscillations, even when none are present. Reanalyzing EEG data from two visual WM experiments (n = 74), and leveraging spectral parameterization, we found systematic changes in aperiodic activity with WM load, and we replicated classic alpha, but not theta, oscillatory effects after controlling for aperiodic changes. Aperiodic activity decreased during WM retention, and further flattened over the occipitoparietal cortex with an increase in WM load. After controlling for these dynamics, aperiodic-adjusted alpha power decreased with increasing WM load. In contrast, aperiodic-adjusted theta power increased during WM retention, but because aperiodic activity reduces more, it falsely appears as though theta “oscillatory” power (e.g., bandpower) is reduced. Furthermore, only a minority of participants (31/74) had a detectable degree of theta oscillations. These results offer a potential resolution to the theta paradox where studies show contrasting power changes. We identify novel aperiodic dynamics during human visual WM that mask the potential role that neural oscillations play in cognition and behavior. Significance statement Working Memory (WM) is our ability to hold information in mind without it being present in our external environment. Years of research focused on oscillatory brain dynamics to discover the mechanisms of WM. Here, we specifically look at oscillatory and non-oscillatory, aperiodic activity as measured with scalp EEG to test their significance in supporting WM. We challenge earlier findings regarding theta oscillations with our analysis approach, while replicating alpha oscillation findings. Furthermore, aperiodic activity is found to be involved in WM, over frontal regions in a task-general manner, and over anterior regions this activity is reduced with an increase the number of items that are remembered. Thus, we have identified novel aperiodic dynamics during human visual WM.


Figure 1) Schematic introducing features of neuro-electrophysiological recordings. A) An example (simulated) time series, with a combination of aperiodic activity and a bursty 10 Hz oscillation. B) The annotated power spectrum for the signal in (A), showing the estimated power of the signal (black) as well as the measured aperiodic component (blue). Frequency ranges are shaded by typical oscillation band ranges -theta: 3-8 Hz, alpha: 8-13 Hz, and beta: 13-35 Hz. C) An example comparison of two power spectra. In this comparison, the difference in the two power spectra was simulated as a change in the aperiodic exponent. D) The quantified parameter differences for the example spectra in (B). When measuring power across pre-defined oscillations bands, there is what appears to be a pattern of changes across bands. However, this can be explained by a change in the aperiodic exponent, which is the parameter that was actually changed in this simulation.
Figure 3) Results Across time. A) Publication years of the literature dataset. Each datapoint represents a 6 month time interval. B-E) Properties of the dataset across time, showing B) number of disorders studied, C) median sample sizes, D) fit methods, comparing spectral parameterization (specparam) and linear regression methods and E) reported motivations and interpretations, reporting if proportion of reports interpreting aperiodic activity as related to E/I ratio and the proportion discussing aperiodic activity as a possible biomarker. Note that in B-E, each time intervals is not an equal length, as papers prior to 2021 are grouped together (due to the low number of papers per year during this time), and the year 2024 including only the first 6 months of the year.
Dataset of all reports investigating aperiodic neural activity with clinical populations
A systematic review of aperiodic neural activity in clinical investigations
  • Preprint
  • File available

October 2024

·

189 Reads

·

2 Citations

In the study of neuro-electrophysiological recordings, aperiodic neural activity – activity with no characteristic frequency – has increasingly become a common feature of study. This interest has rapidly extended to clinical work, with many reports investigating aperiodic activity from patients from a broad range of clinical disorders. This work typically seeks to evaluate aperiodic activity as a putative biomarker relating to diagnosis or treatment response, and/or as a potential marker of underlying physiological activity. There is thus far no clear consensus on if and how aperiodic neural activity relates to clinical disorders, nor on the best practices for how to study it in clinical research. To address this, this systematic literature review, following PRISMA guidelines, examines reports of aperiodic activity in electrophysiological recordings with human patients with psychiatric and/or neurological disorders, finding 143 reports across 35 distinct disorders. Reports within and across disorders are summarized to evaluate current findings and examine what can be learned as pertains to the analysis, interpretations, and overall utility of aperiodic neural activity in clinical investigations. Aperiodic activity is commonly reported to relate to clinical diagnoses, with 31 of 35 disorders reporting a significant effect in diagnostic and/or treatment related studies. However, there is variation in the consistency of results across disorders, with the heterogeneity of patient groups, disease etiologies, and treatment status arising as common themes across different disorders. Overall, the current variability of results, potentially confounding covariates, and limitations in current understanding of aperiodic activity suggests further work is needed before aperiodic activity can be established as a potential biomarker and/or marker of underlying pathological physiology. Finally, a series of recommendations are proposed, based on the findings, limitations, and key discussion topics of the current literature to assist with guiding productive future work on the clinical utility of studying aperiodic neural activity. Project Repository The project repository contains code & data related to this project: https://github.com/TomDonoghue/AperiodicClinical

Download

Figure 3. A functional link across components measured by trial-by-trial Pearson 218 correlations. a. Color-coded values represent the average correlation across participants in the 219 young and old groups in the hippocampus. A significant difference between the young and old 220 groups based on two-sample t-tests is indicated by black boxes. b-c. Bar plots comparing the 221 correlation between theta and spectral slope (b) and between theta and alpha (c) across the 222 young and old participants in the hippocampus. d. Correlation across neural components in the 223 DLPFC. e-f. Bar plots comparing the correlation between spectral slope and theta (e) and 224 between spectral slope and alpha (f) across the young and old participants in the DLPFC. * 225 indicates p<0.05 and *** indicates p<0.001. 226
Figure 4. Neural correlates of age-related spatial memory performance change. a. A 258 schematic for measuring performance by distance accuracy. b. A scatter plot and comparison of 259 the young and older participants by distance accuracy. c-d. A comparison of hippocampal theta 260 power between good and bad performers in each age group (c) and other neural components in 261 the hippocampus (d). e-f. A comparison in DLFPC spectral slope between age groups (e) and 262 other DLPFC neural components (f) across good and bad performers in each age group. 263
Figure 5. Hippocampal dysfunction compensated by DLPFC spectral slope. a. A scatter plot 278 showing a linear relationship between hippocampal theta and distance accuracy defines residual 279 accuracy as the deviance between an actual accuracy and an expected value based on 280 hippocampal theta power. Colored points with the participant number indicate participants who 281 performed better (triangle) or worse (square) than the hippocampal theta-based expectation. b. 282 A scatter plot showing a linear relationship between DLPFC spectral slope and the residual 283 accuracy. Good-performers showed a flatter spectral slope, while bad-performers showed a 284 steeper spectral slope. c. The number of participants after grouping based on hippocampal theta 285 power (top) and DLPFC spectral slope (bottom, only for the "low theta" group). d. Spatial memory 286 performance across the three groups. 287
Figure 6. Impaired view utilization during retrieval in aging and its neural correlates. a-b. 340 Left: A bird-eye view schematic of the virtual environment with an example of encoding and 341 retrieval paths for Match (a) and Nonmatch (b) trials. Right: An example of a view with a distal 342 landmark that participants see during the encoding and retrieval across the two conditions. c. 343 Match -Nonmatch accuracy difference in young and old participants. d. Hippocampal theta power 344 difference between the Match and Nonmatch conditions in young and old participants. e. (top) A 345
Aperiodic neural excitation of the prefrontal cortex offsets age-related decrease in hippocampal theta activity for spatial memory maintenance

October 2024

·

72 Reads

Currently, there is a critical gap in age-related electrophysiological changes in the human brain and how they are correlated with individual decline and maintenance of spatial cognitive function. To characterize these complex neurocognitive changes using direct intracranial recordings, we isolated periodic band power from the aperiodic spectral slope from iEEG power spectra of 69 presurgical epilepsy patients between 19 and 61 years of age while they performed a virtual spatial navigation task. We found a flattening of aperiodic spectral slope in the prefrontal cortex, but also observed a steepening in the hippocampus, suggestive of region-specific changes in excitatory/inhibitory balance across aging. The hippocampus showed pronounced changes in periodic (oscillatory) activity, including a decrease in theta power that correlated with impaired spatial memory, potentially due to changes in the cholinergic system. Interestingly, individuals with the flatter spectral slope in DLPFC showed preserved performance despite lower hippocampal theta, indicating a potential compensatory mechanism for cognitive maintenance. These findings provide new evidence that individual age-related cognitive decline can be predicted by changes in hippocampal theta oscillation, in combination with concomitant prefrontal compensatory mechanisms.


Evaluating and Comparing Measures of Aperiodic Neural Activity

September 2024

·

37 Reads

·

2 Citations

Neuro-electrophysiological recordings contain prominent aperiodic activity – meaning irregular activity, with no characteristic frequency – which has variously been referred to as 1/f (or 1/f-like activity), fractal, or ‘scale-free’ activity. Previous work has established that aperiodic features of neural activity is dynamic and variable, relating (between subjects) to healthy aging and to clinical diagnoses, and also (within subjects) tracking conscious states and behavioral performance. There are, however, a wide variety of conceptual frameworks and associated methods for the analyses and interpretation of aperiodic activity – for example, time domain measures such as the autocorrelation, fractal measures, and/or various complexity and entropy measures, as well as measures of the aperiodic exponent in the frequency domain. There is a lack of clear understanding of how these different measures relate to each other and to what extent they reflect the same or different properties of the data, which makes it difficult to synthesize results across approaches and complicates our overall understanding of the properties, biological significance, and demographic, clinical, and behavioral correlates of aperiodic neural activity. To address this problem, in this project we systematically survey the different approaches for measuring aperiodic neural activity, starting with an automated literature analysis to curate a collection of the most common methods. We then evaluate and compare these methods, using statistically representative time series simulations. In doing so, we establish consistent relationships between the measures, showing that much of what they capture reflects shared variance – though with some notable idiosyncrasies. Broadly, frequency domain methods are more specific to aperiodic features of the data, whereas time domain measures are more impacted by oscillatory activity. We extend this analysis by applying the measures to a series of empirical EEG and iEEG datasets, replicating the simulation results. We conclude by summarizing the relationships between the multiple methods, emphasizing opportunities for re-examining previous findings and for future work.



The Temporal Dynamics of Aperiodic Neural Activity Track Changes in Sleep Architecture

January 2024

·

129 Reads

·

3 Citations

The aperiodic (1/f-like) component of electrophysiological data - whereby power systematically decreases with increasing frequency, as quantified by the aperiodic exponent - has been shown to differentiate sleep stages. Earlier work, however, has typically focused on measuring the aperiodic exponent across a narrow frequency range. In this work, we sought to further investigate aperiodic activity during sleep by extending these analyses across broader frequency ranges and considering alternate model definitions. This included measuring knees in the aperiodic component, which reflect bends in the power spectrum, indicating a change in the exponent. We also sought to evaluate the temporal dynamics of aperiodic activity during sleep. To do so, we analyzed data from two sources: intracranial EEG (iEEG) from 106 epilepsy patients and high-density EEG from 17 healthy individuals, and measured aperiodic activity, explicitly comparing different frequency ranges and model forms. In doing so, we find that fitting broadband aperiodic models and incorporating a knee feature effectively captures sleep-stage-dependent differences in aperiodic activity as well as temporal dynamics that relate to sleep stage transitions and responses to external stimuli. In particular, the knee parameter shows stage-specific variation, suggesting an interpretation of varying timescales across sleep stages. These results demonstrate that examining broader frequency ranges with the more complex aperiodic models reveals novel insights and interpretations for understanding aperiodic neural activity during sleep.




How Can We Differentiate Narrow-Band Oscillations from Aperiodic Activity?

August 2023

·

86 Reads

·

10 Citations

Human intracranial recordings are composed of both periodic “narrow-band oscillations” along with aperiodic, “1/f-like”, activity. While oscillations have been a consistent focus of investigation in neurophysiological data, the physiological and functional properties of aperiodic activity are less well understood, as it is often treated as “background noise”. In this chapter, we provide an overview of both periodic and aperiodic activity, providing a brief historical perspective on the study of each, along with conceptual approaches and analytic assumptions researchers have made when measuring both types of activity. We then highlight recent methodological developments and available techniques for explicitly measuring (a)periodic activity, so as to evaluate which components of the neural data are changing. We propose that studies of human intracranial recordings should employ measures which can differentiate (a)periodic signals in order to determine if they have distinct generators and functional roles. Finally, we discuss putative interpretations of both periodic and aperiodic activity, as well as several unresolved issues which can be explored in future work to further model and interpret these signals.


Figure 2: Canonical spectral power changes. A-B Mean full scalp power spectra for each diagnostic group. Shaded areas represent the standard error of the mean. Cohort1: green = HC, blue = AD; Cohort2: yellow = HC, purple = AD. C-D Comparison of the original SPR computed from the raw power spectra showed a significant difference between the groups in both cohorts. E-F No significant difference was found when considering low frequencies (delta + theta) alone in cohorts 1 and 2. G-H High frequency (alpha + beta) power increased significantly in cohort 1 but did not differ significantly in cohort 2. *** p < .0001, ** p < .001, * p < .05, ns p > .05.
Figure 3: Group averages of parametrized power spectra. A Cohort 1. Mean full scalp power spectra for each diagnostic group after 'specparam' parametrization. The final 'specparam' model fits are in green (HC) and blue (AD). Each power spectrum further consists of periodic activity (shaded area) and the aperiodic component (dashed line). B Cohort 2 (yellow = HC; purple = AD).
Resting-state EEG signatures of Alzheimer's disease are driven by periodic but not aperiodic changes

June 2023

·

213 Reads

·

6 Citations

Electroencephalography (EEG) has shown potential for identifying early-stage biomarkers of neurocognitive dysfunction associated with dementia due to Alzheimer's disease (AD). A large body of evidence shows that, compared to healthy controls (HC), AD is associated with power increases in lower EEG frequencies (delta and theta) and decreases in higher frequencies (alpha and beta), together with slowing of the peak alpha frequency. However, the pathophysiological processes underlying these changes remain unclear. For instance, recent studies have shown that apparent shifts in EEG power from high to low frequencies can be driven either by frequency specific periodic power changes or rather by non-oscillatory (aperiodic) changes in the underlying 1/f slope of the power spectrum. Hence, to clarify the mechanism(s) underlying the EEG alterations associated with AD, it is necessary to account for both periodic and aperiodic characteristics of the EEG signal. Across two independent datasets, we examined whether resting-state EEG changes linked to AD reflect true oscillatory (periodic) changes, changes in the aperiodic (non-oscillatory) signal, or a combination of both. We found strong evidence that the alterations are purely periodic in nature, with decreases in oscillatory power at alpha and beta frequencies (AD < HC) leading to lower (alpha + beta) / (delta + theta) power ratios in AD. Aperiodic EEG features did not differ between AD and HC. By replicating the findings in two cohorts, we provide robust evidence for purely oscillatory pathophysiology in AD and against aperiodic EEG changes. We therefore clarify the alterations underlying the neural dynamics in AD and emphasise the robustness of oscillatory AD signatures, which may further be used as potential prognostic or interventional targets in future clinical investigations.


Citations (25)


... To address these questions, here we analyzed 62 human intracortical recordings obtained during resting state. We systematically studied the relation between frequency and aperiodic exponent using the two complementary analytical methods to estimate aperiodic activity in brain field potentials (Donoghue et al., 2024), namely Specparam, formerly known as FOOOF (Donoghue et al., 2020) and Irregular Resampling Auto-Spectral Analysis (IRASA) (Wen & Liu, 2016). We found that there is indeed a strong positive dependency between frequency and aperiodic exponent. ...

Reference:

Aperiodic exponent of brain field potentials is dependent on the frequency range it is estimated
A systematic review of aperiodic neural activity in clinical investigations

... Aperiodic activity can be examined by measuring the aperiodic exponent (equivalently, the spectral slope) from the neural power spectrum ( Figure 1A-B). Aperiodic neural activity is a dynamic physiological signal, and has been shown to vary systematically through development and in aging (Stanyard et al., 2024;Voytek et al., 2015), across sleep and wake stages (Ameen et al., 2024;Lendner et al., 2020), and during cognitive tasks (Gyurkovics et al., 2022;Waschke et al., 2021). ...

The Temporal Dynamics of Aperiodic Neural Activity Track Changes in Sleep Architecture

... In a sliding window process, lasting one second, 43 metrics were extracted per channel for each study participant ( Table 2 for more details). These metrics included entropy metrics [23][24][25], basic statistical measures [26,27], power spectral density (PSD) metrics [28][29][30], frequency domain metrics [31], as well as fractal dimension and complexity metrics [20,26,32]. After the extraction process, the timeseries data from each channel were compressed using 10 statistical functions: mean, median, minimum, maximum, Journal of Biological Methods | Volume XX | Issue X | standard deviation, variance, and the 25 th , 50 th , 75 th , and 95 th quantiles [33]. ...

Resting-state EEG signatures of Alzheimer's disease are driven by periodic but not aperiodic changes
  • Citing Article
  • December 2023

Neurobiology of Disease

... Oscillatory activities are periodic and exhibit frequency-specific 'narrowband' power, appearing as peaks in the power spectrum. Background activities, on the other hand, are aperiodic and distributed across all frequency bands, typically exhibiting a 1/f power distribution (Barry & De Blasio, 2021;Donoghue & Watrous, 2023). Recent studies have provided compelling evidence that the 1/f component reflects meaningful neural activity, transcending its previous characterization as mere noise (Donoghue et al., 2020). ...

How Can We Differentiate Narrow-Band Oscillations from Aperiodic Activity?
  • Citing Chapter
  • August 2023

... Recent studies have shown that the EEG power spectrum's 1/f-like aperiodic activity partially reflects the overall cortical balance of excitation and inhibition [9,15]. Abnormalities in this measure have been consistently reported [16] and have been linked to cognitive function [17]. Therefore, we used this spontaneous-excitation/inhibition measure to further investigate the relationship between cortical excitability and p-tau181 concentration. ...

Resting-state EEG signatures of Alzheimer's disease are driven by periodic but not aperiodic changes

... The temporal lobe is involved in memory and perception [32], although with a variability in function that has been attributed to single neurons [33], which is suggestive of the capacity to manage many tasks involved in how organisms perceive the external world. Connectivity between the middle temporal gyrus and the hippocampus has been shown to be a key factor in forming new associations between hitherto non-connected entities or concepts [34], and intra-connectivity within the middle temporal gyrus has been linked with social cognition and language [35]. ...

Single neurons in the human medial temporal lobe flexibly shift representations across spatial and memory tasks
  • Citing Article
  • April 2023

Hippocampus

... An open question remains as to how event structure influences our ability to multitask and the consequences for memory. Future research employing multitasking designs holds promise not only for uncovering individual variations in memory performance but also for evaluating the question of whether hippocampal representations are task specific or serve broader, domain general purposes (Han et al., 2023). ...

Using multi-task experiments to test principles of hippocampal function
  • Citing Article
  • April 2023

Hippocampus

... Relatedly, many institutions are now developing introductory data science classes, which teach programming and statistics with broad applicability (Donoghue et al., 2021(Donoghue et al., , 2022Çetinkaya-Rundel and Ellison, 2021). These are powerful courses for neuroscience students to develop intuitions around data and foundational programming skills and can be more accessible (both conceptually and logistically) than introductory computer science courses. ...

Course Materials for Data Science in Practice

Journal of Open Source Education

... As a first example, event-related potentials (ERPs) reflect cross-trial, time-domain average neural responses to internal and external events of interest. There are different kinds of ERP, occurring at various temporal offsets and emerging from different sources in the brain [95][96][97]. A massive body of work has shown that these patterns offer reliable signatures of behavior and internal cognitive processes, including sensory processing [98], prediction error (P300 [99]; N400 [100]), action planning (readiness potential [101]), attention [102], and memory [103]. ...

Automated meta-analysis of the event-related potential (ERP) literature

... Prior studies investigating the EEG power spectra in TSC have largely focused on absolute or relative power. However, parametrization of the power spectrum into aperiodic and periodic components can provide a more accurate estimate of non-oscillatory and oscillatory activity [35,36], and may be valuable in considering differences in excitatory/inhibitory balance and the impact of seizure medications on beta oscillations in TSC (Fig. 1). The aperiodic component of the power spectra is defined by the 1/f power law distribution of the absolute power spectra. ...

Spectral parameterization for studying neurodevelopment: How and why

Developmental Cognitive Neuroscience