Functional analysis within latent states: A novel framework for analysing functional time series data.

Area of Science:

• Neuroscience
• Statistics
• Data Science

Background:

• Functional data analysis (FDA) models data as functions over continua.
• Existing FDA methods may lack interpretability for complex time series.
• Understanding functional brain activity dynamics is crucial for developmental neuroscience.

Purpose of the Study:

• Introduce the FunctionaL Analysis Within LatEnt StateS (flawless) framework for nested FDA.
• Analyze functional time series of electroencephalography (EEG) power spectral densities.
• Investigate latent states of resting-state brain activity in early adolescents.

Main Methods:

• Developed a nested FDA framework (flawless).
• Applied flawless to EEG power spectral density time series from 503 early adolescents.
• Utilized Bayesian regression models to analyze state occupancy dynamics and health associations.

Main Results:

• Identified four distinct functional latent states in adolescent brain activity.
• These states correlate with variations in psychopathology and cognitive function.
• Found significant associations between latent state dynamics, functional traits, and health measures.

Conclusions:

• The flawless framework provides interpretable insights into functional time series.
• It offers a sophisticated approach to analyzing neuroscientific data, reducing assumptions about oscillatory frequencies.
• flawless enhances understanding of longitudinal functional data across research and practice.