Decoding: Negative values output from decoding, availability of v5-topics-50.txt

56 views
Skip to first unread message

Alexis Hervais-Adelman

unread,
May 17, 2019, 9:24:56 AM5/17/19
to neurosynth
Dear Neurosynth developers and users

thanks for the great tool and the already very helpful  discussion on here.

I've been trying to carry out an analysis of the topics most associated with each of a set of maps. Each map is simply a binarised niftii.gz image with 1s in the areas of interest and 0s elsewhere.

I've been following the following procedure:

dataset = Dataset('database.txt')
features = pd.read_csv('features.txt', sep='\t', index_col=0)
dataset.add_features(features, append=False)
decoder = ns.decode.Decoder(dataset,method='roi')
data = decoder.decode('ROI.nii.gz') 

This seems to work, but I would like to be able to understand the outputs. Most confusing to me are the negative values - are these meaningful, or should I threshold at some positive value before thinking about what these mean?

Ultimately, I would like to query a reduced topic space (using the features in the database there are 3228, which is somewhat uninterpretable). The neurosynth website alludes to a 50 topic list (also 400 and 200). Old versions of the 50 topics are available as text files in various places. However, these older versions are not compatible with the current neurosynth database, since the number of studies they contain is lower. This provokes a crash when I try to decode using these other files (e.g. v3-topics-50.txt) as the features.
Therefore, I am looking for v5-topics-50 as a file that I might be able to download and actually use as the features input to the decoder. Does anyone have this, or know how I can generate it? I'm going to admit in advance that handling the gcLDA approach is beyond what I can handle, so I'd be very grateful for either the file or some code that can be used or modified to achieve this.

Thanks very much in advance for any suggestions.
Sincerely
Alexis Hervais-Adelman
Reply all
Reply to author
Forward
0 new messages