Methodological Considerations on the Use of Different Spectral Decomposition Algorithms to Study Hippocampal Rhythms

Area of Science:

• Neuroscience
• Computational Neuroscience
• Signal Processing

Background:

• Local field potential (LFP) oscillations are crucial for neural communication, with specific frequencies implicated in cognitive functions like memory retrieval.
• The "slow gamma" (25-50 Hz) oscillation in the hippocampus is hypothesized to play a distinct role in memory retrieval.
• Theta (θ) harmonics overlap with the slow gamma frequency range, creating ambiguity in identifying distinct oscillatory rhythms.

Purpose of the Study:

• To investigate whether the proposed slow gamma (25-50 Hz) hippocampal oscillations are fundamentally distinct or represent theta harmonics.
• To compare the efficacy of wavelet, ensemble empirical-mode decomposition (EEMD), and Fourier transform methods in resolving these oscillations.
• To critically re-evaluate the definition of fundamental hippocampal oscillations and theories of multiple independent gamma bands.

Main Methods:

• Comparative spectral analysis of local field potential (LFP) data using wavelet transform, ensemble empirical-mode decomposition (EEMD), and Fourier transform.
• Application of these methods to both synthetic datasets containing theta and its harmonics, and real hippocampal data.
• Analysis focused on resolving oscillations within the 25-50 Hz range and distinguishing them from higher-order theta harmonics.

Main Results:

• Ensemble empirical-mode decomposition (EEMD), as previously applied, failed to adequately account for theta harmonics.
• Wavelet analysis, depending on parameters, can erroneously depict slow gamma due to convolution with high-order theta harmonics.
• Fourier decomposition, when applied to high-power events, resolves only theta harmonics in the 25-50 Hz range, challenging the existence of a unique slow gamma rhythm.

Conclusions:

• The slow gamma rhythm (25-50 Hz) may not represent a unique fundamental hippocampal oscillation, but rather theta harmonics.
• Wavelet and EEMD methods can produce misleading results by misinterpreting theta harmonics as slow gamma if not carefully applied.
• The findings necessitate a re-evaluation of the defining characteristics of fundamental hippocampal oscillations and theories based on independent gamma bands.