The HomER program is based on a flexible data architecture, which allows data to be imported with limited preparatory processing from virtually any currently existing commercial NIRS system. The analysis of this data can be tailored to allow the users complete versatility to specify the details of their experiment, including NIRS probe geometry and layout, optical wavelength selection, stimulus design, and other instrumental and acquisition parameters. A screen shot of the HomER program is shown in . Example files and scripts are provided along with the download of the program demonstrating the preconditioning of data for import into HomER. Although HomER is principally designed for CW NIRS measurements, both time-domain [83
] and frequency-domain [136
] data can be analyzed with additional preprocessing. The HomER program has been made freely available and is available from the Photon Migration Imaging Laboratory at the Massachusetts General Hospital ([6
]). This program is available in both its native MATLAB format as open source code and as a compiled standalone executable binary program. The standalone executable is distributed with the MATLAB Runtime Compiler and will run under Windows NT or higher. The open-source MATLAB code is written for MATLAB version 7.0 or higher.
Fig. 7 Screen shot of the HomER program. The layout of the HomER program is based around an interactive graphical display of the NIRS probe, shown in the upper right (b). The user specifies this probe geometry within the data file imported into HomER as described (more ...)
The structure of the HomER program is based on three levels of data processing as shown in . (i) At the first level, individual data files are represented. This level contains the data for a single experimental run, for example, a single acquisition scan. Scan-level processing allows the data from a single experimental run to be interactively visualized, time-series data filtered, and analyzed for functional responses. For a typical experiment, multiple such data files will be recorded within a single experimental session. (ii) The session level represents the second organizing level in HomER. This level includes the joint processing of one or more individual data files for a single experimental subject, recorded during a single session. At the session level, response averages and effects analysis can be calculated from multiple data runs. First-level statistical tests can be performed to investigate the significance of functional changes (i.e., the F test and T test). In addition, images of the changes in optical absorption or hemoglobin concentrations can be reconstructed from these session averages. Session-level data should have consistent optical probe placement and geometry to allow analysis between scans. (iii) At the final level of processing, multisession or group information is represented. This allows group averaging from multiple experimental subjects or multiple sessions from the same subject. Data processing at this level is allowed for region-of-interest comparisons to avoid intersubject registration issues. At all of these three levels, data analysis, visualization, and data export are provided for MATLAB and ASCII formats.
Fig. 8 Levels of analysis in HomER. The HomER program architecture is based on three levels of analysis and processing: a single experimental scan, a session (single subject), and group analysis. At each level, various options for processing, visualization, (more ...)
The primary purpose of the design of the HomER program is to provide the users with easy, quick, and productive interaction and visualization of their data. The program has been designed such that the user should be able to interact with their data from the level of a general overview, handling large data sets, down to the detailed management of individual time courses from a single measurement between a source and detector pair. In HomER, the detailed management of individual measurements is facilitated by a visual layout of the NIRS probe created from user-specified probe geometry and a list of data measurements in the raw data file. By interactively selecting source or detector positions from the NIRS probe, the user is able to toggle the display of different measurements. The time-series data display is color-coded with lines drawn between the corresponding source and detector positions on the NIRS probe. Data can be discarded from analysis, exported to file, or copied to the clipboard function by simply clicking on one of these lines. This probe image allows the user to interact with the data for both the individual data file and session levels of processing. The user can interactively navigate through their data with ease by clicking on the probe geometry exactly as it was set up during data acquisition, allowing the results of an experiment to be easily viewed. The probe also allows users a means to prune measurements from the analysis for subjective reasons, such as the presence of motion artifacts, or objectively based on a low signal-to-noise ratio or source–detector separation criterion.
In addition to the interaction across the spatial dimensions of the probe, HomER allows the user to visually interact with the time-series data by enabling the user to remove subject motion or other artifacts from the data. Regions of time can be highlighted and discarded from functional averaging or deconvolution of the hemodynamic response. These regions can also be discarded from the calculation of the data covariance used in the PCA filtering as described earlier in this paper. In addition to selecting blocks of time, individual stimulus epochs can also be removed prior to functional analysis. These two features allow the user to manually select regions of time that are affected by motion or other artifacts. Several statistical based plots, such as studentized residual analysis for outlier detection, F tests, and other analysis of variance analyses, are also included in HomER for the analysis of functional data.
For filtering, HomER offers both bandpass and PCA-based filtering modules. Bandpass filtering is done by using an finite-impulse response (FIR) model. The default FIR filter used in HomER is a fourth-order Butterworth filter and filters twice in a forward and inverse pass of the data using the MATLAB function filtfilt
. This back-and-forth filter reduces accumulation of phase in the data and offers approximately a twofold better rejection in the stop band. Additional options can be selected to specify the characteristics of the filter model (stop band, order, and type of filter). Three options are given for PCA analysis to allow (i) motion correction (see [64
] for an example), (ii) physiological correction using a separate baseline file (see [117
]), and (iii) physiological correction using components derived and applied to the current file.
In HomER, functional responses can be calculated by using either block averaging or a deconvolution model based on the estimation of an FIR impulse response. These methods use an ordinary least-squares fit of the data. The estimation of the evoked response is performed subsequent to filtering and motion correction. Responses are calculated for both the change in optical density and the change in oxyhemoglobin and deoxyhemoglobin. Individual files (scans) are processed first and then averaged together to calculate the average response for a session. Selected regions of interest can also be averaged within and across subjects. Regions of interest are specified based on either T statistics calculated from the effects of the functional model or from F tests describing the ability of the model to fit the data. Joint probabilities across specific wavelengths or hemoglobin types are not currently calculated.
Finally, HomER also includes basic tools for image reconstruction. HomER currently supports calculations of the optical forward model (sensitivity matrix) from a homogeneous, semi-infinite, slab geometry, using the analytic form given in [11
]. The user is given options to specify the absorption and scattering properties of the medium at each wavelength. Currently only CW forward models are supported. Images are reconstructed by using either the back-projection method (default) or via a Tikhonov regularized inverse. For reconstruction of images of hemoglobin, HomER uses a spectral prior incorporated into the inverse problem as described in [137
]. Movies of the functional response can also be reconstructed and saved by using HomER.