Interscan displacement-induced variance in PET activation data is excluded by a scan-specific attenuation correction

Area of Science:

• Neuroimaging
• Positron Emission Tomography (PET)
• Medical Physics

Background:

• Subject interscan displacements in PET studies can mimic changes in cognitive state by altering regional cerebral blood flow.
• Standard measured attenuation correction (MAC) may not adequately correct for motion-induced artifacts in PET data.
• Artifacts from head motion can confound the interpretation of activation studies.

Purpose of the Study:

• To investigate the impact of interscan displacement artifacts on PET activation studies.
• To assess the efficacy of scan-specific calculated attenuation correction (CAC) versus MAC in removing these artifacts.
• To determine if CAC can prevent variance induced by interscan displacements.

Main Methods:

• Analysis of two independent PET datasets with varying degrees of interscan displacement (large and small).
• Application of both calculated attenuation correction (CAC) and measured attenuation correction (MAC).
• Utilized SPM99 for data analysis, incorporating interscan displacement parameters (IDP) as covariates in the General Linear Model.

Main Results:

• For large displacements, CAC dramatically reduced the affected brain volume by approximately 50-fold compared to MAC.
• Even with small displacements, CAC significantly reduced the affected brain volume approximately 10-fold compared to MAC.
• Introducing IDP covariates with MAC did not effectively resolve the variance caused by interscan displacements.

Conclusions:

• Scan-specific calculated attenuation correction (CAC) is a superior method for preventing interscan displacement-induced variance in PET studies.
• CAC is more effective than MAC, even when IDP covariates are used with MAC, particularly for minimizing motion artifacts.
• Applying CAC is recommended for all PET studies, especially those with even minor interscan displacements, to ensure data integrity.