Selecting causal brain features with a single conditional independence test per feature
Authors: Atalanti Mastakouri, Bernhard Schölkopf, Dominik Janzing
NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate successful application of our method to simulated, as well as encephalographic data of twenty-one participants, recorded in Max Planck Institute for intelligent Systems. The detected causes of motor performance are in accordance with the latest consensus about the neurophysiological pathways, and can provide new insights into personalised brain stimulation. |
| Researcher Affiliation | Collaboration | Atalanti A. Mastakouri Empirical Inference Department Max Planck Institute for Intelligent Systems Tübingen, 72076 amastakouri@tue.mpg.de Bernhard Schölkopf Empirical Inference Department Max Planck Institute for Intelligent Systems Tübingen, 72076 bs@tue.mpg.de Dominik Janzing Amazon Research Tübingen, 72076 janzind@amazon.com |
| Pseudocode | Yes | Algorithm 1: Find causes of R Input: P i, M i, R, i = 1, ..., n. Output: Causes R for i 1 to n do pvalue1 ind_test(M i, R) if pvalue1 < threshold1 then pvalue2 cond_ind_test(P i, R, M i) if pvalue2 > threshold2 then Causes R [Causes R, M i] end end end |
| Open Source Code | Yes | The Matlab code can be found in https://gitlab.tuebingen.mpg.de/amastakouri/single CICausal Feature Selection.git |
| Open Datasets | No | The paper states 'We apply our method on simulated data and on EEG data that we recorded from twenty-one healthy participants.' and 'We recorded twenty-one healthy participants with high density EEG (128 electrodes, Brain Products), during a motor task.' It does not provide any concrete access information (link, DOI, specific repository, or citation to a public version) for this dataset. |
| Dataset Splits | No | The paper does not explicitly provide specific training/validation/test dataset splits (e.g., percentages, sample counts, or references to predefined splits). |
| Hardware Specification | No | The paper states 'Simulations were run on a 12-CPU computer using the parallel toolbox of Matlab.' but does not specify the exact CPU model or other hardware details like GPU or memory. |
| Software Dependencies | No | The paper mentions 'Matlab code' and specific tests like 'HSIC test [23]' and 'conditional independence HSIC test from [24, 25]' but does not provide version numbers for Matlab or any other specific software dependencies. |
| Experiment Setup | Yes | For the statistical testing we examine the null hypothesis H01 : M i R and consider to have rejected the null hypothesis (hence consider to have found M i and R to be dependent) if p < αD = 0.05. Then, we examine the null hypothesis H02 : P i R | M i and accept it (hence the conditional independence) if p > αCI = 0.25 (usual values for accepting CI in EEG datasets include thresholds above 0.25 [8]). |