New Insights into Signed Path Coefficient Granger Causality Analysis.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Econometrics

Background:

• Granger causality is a time series analysis technique applied across neuroscience modalities like fMRI, EEG/MEG, and fNIRS.
• The "signed path coefficient Granger causality" method, derived from econometrics, has gained traction in recent fMRI research.
• This method interprets autoregressive coefficients as "excitatory" or "inhibitory" influences.

Purpose of the Study:

• To evaluate the validity of the "signed path coefficient Granger causality" method in neuroscience research.
• To determine if the autoregressive coefficients accurately reflect true causal relationships in time series data.
• To identify potential limitations and erroneous conclusions arising from the application of this method to fMRI data.

Main Methods:

• Conducted computational analyses using resting-state fMRI data.
• Performed simulation experiments to rigorously test the signed path coefficient Granger causality method.
• Examined the consistency between autoregressive coefficients and actual causal relationships.

Main Results:

• Demonstrated that the signed path coefficient Granger causality method is flawed and untenable.
• Found that autoregressive coefficients do not consistently align with real causal relationships.
• Highlighted the potential for significant misinterpretations due to erroneous conclusions.

Conclusions:

• The applicability of signed path coefficient Granger causality in fMRI research is severely limited.
• Researchers must exercise caution when employing this method to avoid misinterpreting causal influences in neuroimaging data.
• Further validation and development of causality analysis techniques for neuroscientific applications are warranted.