2021年1月22日　　Friday Lunch Seminar
12:15 〜 13:00

(CiNet メンバーのみを対象に On-lineで開催いたします。事前申し込み要）

演題：　“i-test: alternative to t-test for the decoding accuracy evaluation”

情報通信研究機構(NICT)
脳情報通信融合研究センター(CiNet)
脳情報通信融合研究室　研究員
廣瀬 智士

担当PI : 内藤 栄一

Abstract:
In fMRI decoding studies we often test the mean decoding accuracy across participants against the chance-level accuracy (e.g., one-sample Student t-test or permutation test) to verify whether brain activation includes information about the label (e.g., experimental condition, cognitive content). However, the significant results for such tests only indicate that “there are some people in the population whose fMRI data carry information about the experimental condition.” (Allefeld et al., 2016) Thus, such tests failed to infer whether the effect is typical in the population. In this study, I propose the statistical test “information prevalence inference using the i-th order statistic (i-test).” The i-test has high statistical power to provide an inference regarding the typical effect in the population. In the i-test, the i-th lowest sample decoding accuracy (the i-th order statistic) is compared to the null distribution to test whether the proportion of higher-than-chance decoding accuracy in the population (information prevalence) is larger than the threshold. Thus, a significant result of the i-test can be interpreted as “a majority of the population has information about the label in the brain.”

Allefeld, C., Görgen, K., Haynes, J.D., 2016. Valid population inference for information-based imaging: From the second-level t-test to prevalence inference. Neuroimage