Here are some statistics and analyses of the neuroimaging datasets, as well as annotations on each movie. Below you will find an interactive map of the Cohen's d effect size for tSNR. We compared tSNR for raw data ('raw'), minimally-preprocessed data up to blurring ('blur') and fully-preprocessed data ('pre').

Initial value
    Color palette:


    Pos. threshold:
    Neg. threshold:

    Movie List

    This table provides information on the movies screened for the NNDb project. We aim to expand this selection over time and make the NNDb even more comprehensive.

    Movie Year Length (sec) Genre IMDb Score
    12 Years a Slave 2013 7715 Historical 8.1
    500 Days of Summer 2009 5470 Romance 7.7
    Back to the Future 1985 6674 Sci-fi 8.5
    Citizenfour 2014 6804 Documentary 8.1
    Little Miss Sunshine 2006 5900 Comedy 7.8
    Pulp Fiction 1994 8882 Action 8.9
    Split 2016 6739 Horror 7.3
    The Prestige 2006 7515 Thriller 8.5
    The Shawshank Redemption 1994 8181 Drama 9.3
    The Usual Suspects 1995 6102 Crime 8.6

    Data Quality Statistics

    Here you will find stats on data quality: the table shows percentage of manually identified independent component analysis (ICA) artifacts out of 250 dimensions, and statistics on the raw (Raw), blurred (Blur) and fully detrended and pre-processed (Pre) timeseries data by comparing temporal signal-to-noise ratios (tSNR).

    Measure ICA Artifacts (%) tSNR
    Mean Max
    Raw Blur Pre Raw Blur Pre
    Min 56.40 11.85 12.20 13.37 91.33 110.33 185.85
    Mean 71.71 39.43 44.55 63.82 161.19 201.20 319.18
    SD 7.47 10.17 12.73 20.79 29.15 40.16 50.58
    Max 87.20 60.10 68.99 98.03 218.09 310.91 431.79

    Stimulus Annotations

    Below are annotations of faces and speech for the movies screened. Click on each movie poster for seeing a table of annotations, and for downloading the data. If you would like to contribute to annotating movies, or you are interested in machine learning competitions, please do get in touch with or, or head over to our Github page.