Skip to content

Latest commit

 

History

History
43 lines (35 loc) · 2.51 KB

README.md

File metadata and controls

43 lines (35 loc) · 2.51 KB
Raw

nirskit

About

nirskit is a wrapper for the NIRS Brain AnalyzIR toolbox. It was designed with the objective of trading some flexibility for ease-of-use.

The main design principles are:

  • Each function should map onto a cohesive stage of the pipeline (preprocessing, subject-level, group-level, etc)
  • Mandatory steps (e.g., optical density conversion, Beer-Lambert law) should be done automatically
  • Analysis parameters should be passed in as key-value pairs
  • Functions should use directories as inputs and outputs
  • The processing pipeline should be encoded in the generated file paths

Usage

This design allows much faster and easier execution of typical pipelines. For example:

% Preprocess with TDDR motion correction, resample to 5Hz, and high-pass filter with a cutoff of .01 Hz
nirskit.analysis_2_preprocess('MotionCorrection','TDDR','Resample',5, 'HPF',.01);
% Subject-level activation with estimation of temporal/dispersion derivatives, and HRF peak at 6 seconds
nirskit.analysis_3_activation_indiv('InclDeriv',true,'PeakTime',6);
% Group-level activation treating condition as a fixed effect and subject ID as a random effect
nirskit.analysis_4_activation_group('Formula','beta~-1+cond+(1|subject)');

In addition, it simplifies the task of varying a parameter over a range to test its effect:

% Preprocess with high-pass filter cutoffs varying from .001 Hz to .5 Hz, in intervals of .001 Hz
for cutoff=.001:.001:.5
    nirskit.analysis_2_preprocess('HPF',cutoff);
end

Output

Each function takes in a directory that contains a Results.mat file for its input. The expected contents of this file depends on the function (e.g., preprocessing expects raw and emits hb, subject-level activation expects hb and emits SubjStats, group-level activation expects SubjStats and emits GroupModel). The output is then written to a new directory within the input directory, with the directory name specified by the function and its configuration parameters. Therefore the path to any given Results.mat file will reflect the entire processing pipeline used. For example, preprocessed data in the example above would end up in a location like this:

<datadir>\2_preprocessed_MotionCorrection-TDDR_Resample-5Hz_HPF-0.01Hz\Results.mat

And the corresponding group-level results would end up in a location like this:

<datadir>\2_preprocessed_MotionCorrection-TDDR_Resample-5Hz_HPF-0.01Hz\3_activation-indiv_conditions_Peak-6s-deriv\4_activation-group_beta~-1+cond+(1|Name)\Results.mat