Fixed bootstrap tests and included BCa and percentile confidence intervals #24
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…es without addressing.
… Need to refer to their Number properties.
…lier versions it was a warning. In 2014a it is an error.
Move to package structure
Conflicts: +rsa/+fig/pageFigure.m
Function getUserOptions let’s you specify input parameters flexibiy as structures (with fields) or in the ‘fieldname’, value, … style. Should be used consistently in the toolbox.
Utility function to deal with User Options flexibily
different. Flexible transform functions are allowed (not just rank transform)
OLS model estimation and MANOVA model
Added the searchlight toolbox. Functionality added for running volumetric and surface based searchlights. Also added functionality for reading and writing gift files.
Image RDM can now work on an array of structs or on a data frame
Searchlight toolbox
…pairwise comparisons.
1. Speeding up covdiag 2. Additional (optional) input parameter to covdiag (df) 3. Options to do ‘runwise’ and ‘overall’ noise normalisation in distanceLDCraw and noiseNormalizeBeta 4. replaced Sw^-0.5 with inversion over eigenvalues, making the code ~20% faster and numerically (potentially) more stable.
to include former distanceLDC. Otherwise more universal handling of input arguments.
WARNING - these need to be still beta-tested!
pairMatrix is like indentityMatrix(‘allpairs’), but much faster
Small changes for SPM12 compatibility.
… 2016) noiseNormalizeBeta returns also optional trace(R*R’) covdiag takes optimal parameter to fix shrinkage/ varianceLDC does the variance prediction
runSearchlightLDC now uses the design structure to obtain optimal estimates of the LDC distance
Matlab r2104a compatibility - code reviewed, fixes work.
The value of `y` was not being set in case there were no significant …
This file has existed outside the github repository for years, I'm just adding it in.
Include missing exhaustivePermutations file.
- Structure containing Mask image to rsa.defineSearchlight_surface now requires the mask variable (was previous data variable) - Moved readMask to base directory, now is rsa.readMask and can be used to read mask images in the correct format for rsa.defineSearchlight_surface - Moved readSurf to base directory, now is rsa.readSurf - Added surfing functions in +surf package (required for rsa.defineSearchlight_surface) - Fixed calling function to surfing_coords2linvoxelidxs in rsa.defineSearchlight_surface so that Mask is applied when defining searchlights - Modified documentation for a handful of searchlight functions
- Readme.md contains overview of toolbox - Manual in markdown is in rsa_toolbox_manual.md - info.txt contains a link to a primer on using markdown to write formatted text - makepdf.txt contains information on how to compile markdown manual file to create a pdf
…2candRDMs.m, made sure that FDRthreshold doesn't return empty p if nothing gets significant
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Dear RSA group,
I have fixed the bootstrap tests, which had been implemented incorrectly in the previous version of the RSA toolbox. The new function rsa/stat/bootConfint now calculates confidence intervals (default 95% CIs) using two different methods: bias-corrected and accelerated CIs or CIs based on the percentile method (see Efron & Tibshirani, 1993). These CIs are used for inference on relatedness between candidate and reference RDMs (one-tailed) and comparison between candidate RDM correlations (two-tailed test).
The bootstrapping procedure itself is now changed so that on each iteration the statistic of interest (mean correlation) is calculated in the same way as in the original sample to yield valid bootstrap distributions.
Minor fix: FDR-based significance thresholds cannot be "empty" anymore.
The code has been tested with DEMO1 while switching from signed rank test to the bootstrap tests.
I hope that you find these changes useful.
Best,
Michael