MVPA and the Representations of Faces and Objects in Ventral Temporal Cortex
MS Al-Rawi
Hi ...
I am getting ~80% classification accuracy (total_perfs) using {mvpa-log_reg, 577 voxels, subject example of MVPA which is one of the subjects used by Haxby et al} compared to ~83% that was published by Haxby et al (Science, 293, pp 2425-2430, 2001) for the region {All ventral temporal object-selective cortex}. It must be noted that Haxby et al determined the stimulus category that a subject was viewing by using the correlation to measure the similarity between the patterns of response evoked by each category on even and odd runs.
In Haxby et al, accuracies for identifying patterns for individual categories as compared with all other categories ranged from 100% (faces, houses, and scrambled pictures) to 67% (bottles and scissors) these are the lowest.
Using MVPA-log_reg on one of Haxby et al subject (misses two runs), accuracies were 0.88 0.97 0.80 0.30 0.87 0.77 0.78 0.92 0.81 for (faces, houses, cats, bottles, scissors, shoes, chairs, scrambled pictures) respectively. Thus we never get 100% accuracy, and cats gave 30%....!?
Regards
Al-Rawi
Greg Detre
I think that Jim was averaging all the timepoints in a block together
for the analyses he reported, wasn't he? I haven't looked back at the
original paper in a little while though, so correct me if I'm wrong.
g
>
> You received this message because you are subscribed to the Google
> Groups "Princeton MVPA Toolbox for Matlab" group.
> To post to this group, send email to mvpa-t...@googlegroups.com.
> To unsubscribe from this group, send email to
> mvpa-toolbox...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/mvpa-toolbox?hl=en.
--
---
Greg Detre
cell: 617 642 3902
email: gr...@gregdetre.co.uk
web: http://www.princeton.edu/~gdetre/