We run many tests on the raw data captured by the 320 × 256 HgCdTe cooled thermal imager, and make a comparison between our approach with the previous method of bilateral filtering digital detail enhancement and guided image filtering digital detail enhancement.
Differences in the age of participants as well as observations of cognitive heterogeneity in CP participants [24] prompted us to resort to a threefold statistical analysis: First we tested for group differences in location using a robust, non-parametric Wilcoxon rank-sum test on the raw data (average values - mean - for each participant).
Kolmogorov-Smirnov tests on the raw reaction times (RT) showed that the data within the two conditions were normally distributed (all ps>.11).
Statistical analyses were performed using a two-tailed Student's paired t-test on the raw data with GraphPad Prism 4 software (GraphPad Software, La Jolla, CA, USA).
This explains why the change trial detected by this method was significantly earlier than the change trials detected by the application of the change point test on the raw or the filtered firing rates.
We based our association tests on the raw intensities in order to avoid the known loss of information and statistical power resulting from classifying raw copy number measurements into discrete classes, i.e. CNV calls (36, 37).
Significance of differences was determined with a two-tailed Fisher's exact test on the raw data (NSS and SS) with the online calculator available at http://faculty.vassar.edu/lowry/VassarStats.html (accessed October 22, 2009) and a McDonald-Kreitman test [ 66] conducted in DnaSP v. 5.10.
Maxima for final double event increments lie at 7.75 nm and 7.73 nm for iLOV-Coh Doc (wiLOV-Coh DocGFP (N = 255) and Xyn-Doc (GS)8:Coh-CBM (N = 320) final ruptures, respectively (a twild-type -sfGFPt on the rawild-type -sfGFP no significant difference in the data distributions (p-value of 21.7%).
In both sites these changes were supported by the results of a matched t test based on the raw scores, rostral area 35 t(7) = 2.24, p =.031; rostral area 36 t(7) = 1.93, p =.048, one-tailed.
The ROC curve illustrates the rate of true positives as a function of the rate of false positives (when testing different thresholds on the raw SVM output; see Methods for output function).
