Print Close

Novel Cyclic Homogeneous Oscillation Detection Method for Neural Oscillatory Mapping

Poster No:

1628

Submission Type:

Abstract Submission

Authors:

Hohyun Cho¹^,2, Markus Adamek¹^,2, Jon Willie¹^,2, Peter Brunner¹^,2

Institutions:

¹Washington University School of Medicine in St. Louis, St. Louis, MO, ²National Center for Adaptive Neurotechnologies, St. Louis, MO

First Author:

Hohyun Cho
Washington University School of Medicine in St. Louis|National Center for Adaptive Neurotechnologies
St. Louis, MO|St. Louis, MO

Co-Author(s):

Markus Adamek
Washington University School of Medicine in St. Louis|National Center for Adaptive Neurotechnologies
St. Louis, MO|St. Louis, MO

Jon Willie
Washington University School of Medicine in St. Louis|National Center for Adaptive Neurotechnologies
St. Louis, MO|St. Louis, MO

Peter Brunner, PhD
Washington University School of Medicine in St. Louis|National Center for Adaptive Neurotechnologies
St. Louis, MO|St. Louis, MO

Introduction:

Temporal and spectral nuances of neural oscillations are pivotal for decoding complex rhythms underlying large-scale brain activity. Traditional methods, which discern these oscillations through frequency peaks above 1/f-sloped background activity, are often constrained to the frequency domain and fail to precisely determine the onset and offset of oscillations. They cannot distinguish the fundamental frequencies of non-sinusoidal oscillations from their harmonics.

Methods:

Our novel cyclic homogeneous oscillation detection method (CHO) distinguishes oscillations as peaks above the 1/f-sloped noise across time and frequency domains. CHO requires neural oscillations to complete at least two full cycles to ensure their validity and rejects false positives commonly generated by non-sinusoidal waveforms and their harmonics.

·Figure 1 The algorithm of CHO (A) and Performance of CHO in detecting synthetic non-sinusoidal oscillations (B).

Results:

We verified the ability of CHO to isolate simulated oscillatory patterns, both sinusoidal and non-sinusoidal, amidst the prevalent 1/f noise. We applied CHO on a wide range of electrophysiological datasets, including ECoG, EEG, and SEEG recordings, across different cognitive states and tasks. For each of these datasets, CHO accurately detected peak frequencies in task-relevant brain regions, identifying oscillations of specific fundamental frequency in areas such as primary auditory cortex, primary motor cortex, Broca's area, and the hippocampus, both during rest and task engagement. CHO identified characteristic oscillations in ECoG and EEG signals in an auditory reaction-time task; and detected a significant decrease in the initiation of 7 Hz oscillations in the hippocampus, during a memory task.

·Figure 2 Application of CHO in determining the fundamental frequency and duration of neural oscillations during resting state (A-F) and during a recognition memory task (G and H).

Conclusions:

In summary, our results show that CHO stands out as an exceptionally precise and specific instrument for neural oscillation detection, and that it can be used to elucidate the spectro-temporal dynamics of neural oscillations that govern human cognition. CHO considers non-sinusoidal oscillation attributes, such as waveform asymmetry and shape. With this, CHO has the potential to become an important tool in investigating the role of neural oscillations within the brain's circuitry and identifying unique oscillatory biomarkers that may indicate unusual brain functions.

Modeling and Analysis Methods:

EEG/MEG Modeling and Analysis ¹

Methods Development ²

Keywords:

Data analysis

ELECTROCORTICOGRAPHY

Electroencephaolography (EEG)

ELECTROPHYSIOLOGY

Other - Neural Oscillation

^1|2Indicates the priority used for review

Provide references using author date format

Annas, G.J. (1997a), 'New Drugs for Acute Respiratory Distress Syndrome', New England Journal of Medicine, vol. 337, no. 6, pp. 435-439