DREAM

DREAM decodes neural oscillations across multiple frequency bands to enable frequency-specific analysis of brain rhythms and associated behaviors.

Key Features:

• Frequency decomposition: Decomposes neural oscillations into discrete frequency intervals linked to specific physiological processes.
• Log-linear segmentation: Uses the natural log-linear distribution of neural oscillations to partition frequencies into bands.
• Sampling-aware interval calculation: Computes decodable frequency intervals from the sampling period and the number of samples in time-series recordings.
• Time-series modality support: Operates on time-series signals including head motion traces and fast functional MRI data.
• Multi-band amplitude mapping: Quantifies oscillation amplitudes across multiple frequency bands.
• Reliability assessment: Evaluates measurement reliability and inter-individual variability of band-specific amplitude measures.

Scientific Applications:

• Developmental head motion analysis: Applied to head motion oscillations in children, revealing that younger children show greater movement across frequency intervals and that boys aged 7–9 years exhibited greater motion than girls, with higher-frequency bands linked to increased movement and stronger age correlations but weaker sex interactions.
• Fast fMRI analysis (Human Connectome Project): Mapped neural oscillation amplitudes across multiple frequency bands and found higher-frequency bands produced more reliable amplitude measurements, indicative of greater inter-individual variability.
• Frequency-specific brain mapping: Enabled frequency-specific mapping of brain rhythms and assessment of measurement validity across bands.

Methodology:

Segments time-series signals into discrete frequency bands using a natural log-linear frequency partition and computes decodable frequency intervals based on sampling period and sample count, then quantifies band-specific amplitudes and reliability.