Haptic fMRI: using classification to quantify task-correlated noise during goal-directed reaching motions

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Cognitive Neuroscience

Background:

• Neuroimaging artifacts in Haptic fMRI can lead to false positive or correlated activation signals.
• Distinguishing true neural activity from motion artifacts is crucial for reliable Haptic fMRI studies.

Purpose of the Study:

• To investigate whether 3D goal-directed reaching motions using a Haptic fMRI Interface (HFI) introduce confounding motion artifacts.
• To quantify and localize potential artifacts generated by the HFI during fMRI scanning.

Main Methods:

• Simultaneous fMRI scanning of a subject's brain and a customized soft phantom near the motor cortex.
• Utilizing a linear max-margin classifier to differentiate between motion planning and reaching tasks using fMRI time series data.
• Scanning a stand-alone fBIRN phantom with the HFI outside the scanner to isolate interface-specific artifacts.

Main Results:

• fMRI measurements in the phantom were uninformative for task classification (45-73% accuracy).
• Brain regions including primary motor, visual, and somatosensory cortex accurately classified task conditions (90-96% accuracy).
• Artifacts from the haptic interface alone showed lower temporal noise compared to brain phantom measurements.

Conclusions:

• Performing 3D reaching motions with the HFI does not create significant confounding motion artifacts in fMRI.
• Any observed fMRI artifacts in Haptic fMRI reaching experiments are likely dominated by genuine neural responses.
• The study validates the use of Haptic fMRI for investigating motor control without introducing substantial artifactual signals.