Voltage-based versus factor score-based source localization analyses of electrophysiological brain activity: a...

Area of Science:

• Neuroscience
• Computational Neuroscience
• Biomedical Engineering

Background:

• Electrophysiological recordings traditionally offer limited spatial information on neural activity.
• Advancements in inverse problem algorithms are enabling access to spatial origins of neural activation.
• Principal Component Analysis (PCA) can extract component-specific information from event-related potentials (ERPs).

Purpose of the Study:

• To investigate a novel strategy for improving the reliability of neural source localization solutions.
• To compare the effectiveness of applying inverse problem algorithms to factor scores versus traditional voltage measures.
• To determine if using PCA-derived factor scores enhances the salience of neural activation foci.

Main Methods:

• Applied the LORETA algorithm for source localization to motor potentials recorded from 25 subjects.
• Compared LORETA solutions derived from peak voltage, average voltage, and PCA-derived factor scores.
• Utilized a button-pressing task to elicit a well-characterized motor potential.

Main Results:

• All three modalities (peak voltage, average voltage, factor scores) identified the contralateral motor cortex as the primary source of the motor potential.
• Source localization based on factor scores yielded a "cleaner" solution, with the main focus being more salient.
• Factor scores reduced the influence of secondary, potentially noisy, activation foci compared to voltage-based methods.

Conclusions:

• Applying inverse problem algorithms to PCA factor scores represents a more reliable strategy for neural source localization.
• Factor scores provide a "clean amplitude" that enhances the clarity and reliability of identified neural activation origins.
• This approach improves the spatial resolution and interpretability of electrophysiological data.