todo.txt

== Project Management ==

 * more specialized tutorials, e.g. artifact removal, binning.

 * version management - procedures for testing new releases, when to
   create a new version branch, where/how to publish new versions for
   public use, etc.

 * make clearer separation between core and EEG functions

== General Issues ==

 * work toward greater compatibility with FieldTrip. Integrating cluster
   statistics will be a good be a proof-of-concept that will demonstrate this
   potential. Could even approach FieldTrip developers about using
   eeg_ana as a batch handler, leading to a much wider audience. This would
   require a great deal of compatibility work, since FieldTrip supports more
   data types than eeg_ana.

 * improve consistency:

  * naming conventions - getobj vs. get_mat

  * parameter names - chanFilter vs. chanbins vs. eog_channels

  * remove "basic" directory, move all "obj" functions to a "core" directory

  * expand unit tests; pattern modification operations especially
    important

 * check feasibility of cross-experiment analyses (e.g. running Per's
   SME-serial position interaction analysis on all LTP FR experiments)

  * possibility: expand merge_exps.m, add 'experiment' field to each
    subj (or session) - could also add functions for running analyses at
    the session level

== Basic ==

 * make an init_subj function; currently get_sessdirs is the only
 programmatic way to make a subj structure

 * remove init_scalp and init_iEEG, expand import_channels

== Events ==

 * tutorial on use of event filters, binning

== Patclass ==

 * integrate load_pattern features into get_mat, replace load_pattern
 with get_mat in all toolbox functions

 * expand feature selection capability

 * support for importance maps

== Patterns ==

 * support for continuous patterns - one event, many time bins?  How is
   events information stored?

 * move to using EEGLAB-style chan structs

== Plotting ==

 * rather than diff logical, specify linear combination weights

 * option to use locations from the chan struct, rather than from
   file. Would allow topo plots with a subset of channels

 * support for making brain plots

== Stats ==

 * proposal for compatibility with plotting functions

  * move multiple comparisons correction, ect. to functions that operates on
    stat objects

  * standard variable names: p, statistic, thresh, h

  * can we have some sort of setter function for standard stat
    variables, or just rely on help pages to instruct user on creating a
    compatible stat object?

  * thresh may be scalar or matrix (allows for levels of multiple
    comparisons correction within pattern)

  * 1) plot pattern 2) plot pattern with markers 3) plot statistic 4)
    plot thresholded statistic

  * color map is simply statistic, with values below thresh in white

  * translating p to z, setting z* can be done outside of plotting
    function

 * settle "stat" vs. "stats" controversy

 * support for stat printing in exp_disp

 * bootstrapping/summed z

 * RMAOV; expand to N-way, support importing back into MATLAB

 * permutation test on subject significance

 * cluster statistic; integrate ZDC's functions