Whereas a strength of functional MRI is its spatial resolution (or more accurately, the ability to directly solve the MRI image reconstruction problem as compared with the ill-posed nature of spatial reconstruction in diffuse optical methods), strengths of optical methods are certainly temporal resolution and the ability to measure both hemoglobin species more directly. Sophisticated time-series analysis of functional optical studies is thus very important for interpreting such studies. Although this topic is the focus of the remainder of this report, we will begin by acknowledging that much work is still needed in this field. In particular, the analysis of optical measurements poses several unique challenges.
Because fMRI and optical experiments usually have similar designs and hypotheses, many analysis approaches suited for fMRI have often been applied with little modification to optical data [94 (link)]. Indeed, much of the analysis of optical imaging has benefited from similar advances in fMRI. However, because of the different biophysics associated with the optical and fMRI techniques, there are a number of specific issues, limitations, and methods specific to the optical technology. Unlike fMRI analysis, which often draws statistical information from the spatial proximity of measurements (e.g., voxels in the volume images) and temporal compression methods from the assumption of canonical temporal shapes of the evoked response (e.g., the Γ function response [95 ]), analysis of optical data has generally focused on more traditional time-series methods, including bandpass filtering, temporal smoothing, and linear deconvolution, to preserve temporal information about the evoked functional hemodynamic response while trying to remove specific artifacts in the measurements, such as cardiac pulsation signals. In the following sections, we discuss the analysis of the temporal information in optical measurements and emphasize the differences in comparison with fMRI data analysis.
Because fMRI and optical experiments usually have similar designs and hypotheses, many analysis approaches suited for fMRI have often been applied with little modification to optical data [94 (link)]. Indeed, much of the analysis of optical imaging has benefited from similar advances in fMRI. However, because of the different biophysics associated with the optical and fMRI techniques, there are a number of specific issues, limitations, and methods specific to the optical technology. Unlike fMRI analysis, which often draws statistical information from the spatial proximity of measurements (e.g., voxels in the volume images) and temporal compression methods from the assumption of canonical temporal shapes of the evoked response (e.g., the Γ function response [95 ]), analysis of optical data has generally focused on more traditional time-series methods, including bandpass filtering, temporal smoothing, and linear deconvolution, to preserve temporal information about the evoked functional hemodynamic response while trying to remove specific artifacts in the measurements, such as cardiac pulsation signals. In the following sections, we discuss the analysis of the temporal information in optical measurements and emphasize the differences in comparison with fMRI data analysis.