Index of /MISC/ReadingGroup.11/Christoph.12/SCO

This is a description of the items in the SCO folder:
1. Data files- The MS Access files named "Data_(name of subject)" are the data tables having the data for each subject. The data for two subjects, Lucas and Savannah, are also included even though they were not used for the analysis.For each file, N is the number of presentations for each consonant or vowel, while the p is the number of correct responses. The symbols are
symbols = {'@', 'A', 'C', 'D', 'E', 'G', 'I', 'J', 'O', 'R', 'S', 'T', 'U',...
    'W', 'Y', 'Z', 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'k', 'l',...
    'm', 'n', 'o', 'p', 'r', 's', 't', 'u', 'v', 'w', 'y', 'z'}.

For N and p matrices, the first dimension is whether the sound was played in the initial or final position. the second dimension is the target sound (oddball), and the third dimension is the masking sound.

For the CV syllables, one of two phones is kept the same, while the other is changed. target 1 is the consonant or vowel in the initial, while target 2 is the final position. Likewise, mask 1 is the initial pos, and mask 2 is the final pos. The score is 1 when the child gets it correctly, and 0 when wrong obviously. Finally, var_pos states whether the initial or final position is changed. 1 for initial pos and 2 for final pos. 

2. CM- This folder contains excel sheets that represent the confusion matrices for each subject. Each excel sheet has 5 sheets. Sheet 1 is the CM for vowels presented in the initial position, sheet 2 is the CM for vowels presented in final position, sheet 3 is the CM for consonants presented in the initial position, sheet 4 is the CM for consonants presented in the final positon and sheet 5 is the CM overall irrespective of position or consnant/vowel.

For each sheet, the rows are the maskers and the columns are the targets. Blanks mean that there were no presentations and for the cases where there were presentations, the number of times the target was reported CORRECTLY in the presence of the masker is given with the score.i.e. 1/3=0.333, means of the 3 times the masker and target combination were presented, the subject correctly identified the target only once and so the overall score for the combination was 0.333.

The file matrix.m was used to generate this.

3. Scatter plots- These basically are the pdf files scatcer and scatver. I have also provided the matlab figure files of the same name, in case an edit is needed. The file scatcer shows a scatter plot of the error of consonant in the initial position versus consonants in the final position and scatver do the same for vowels. None of these are percentage errors. Every time a target consonant or vowel was not identified as the oddball, that is counted as an error. The errors here are averaged across all maskers for each of the consonants and vowels. The disadvantage of this plot is that since the number of presentations for the SCO task were very different, there is no way to know which subject got how many presentations and how many for each sound. This is just an average measure.

The file exelnscat.m was used to generate this.

4. SCO overall data- The SCO data is summarized in SCO.xlsx. This has eight sheets, two each for consonants in initial position, consonants in final position, vowels in initial position and vowels in final position. The first sheet for each shows the rounded off percentage errors for each subject for all the target sounds, while the second sheet gives a description of how many presentations there were for each sound and how many the subject answered incorrectly. An error is counted when a subject incorrecly identifies the oddball in the three intervals played. For each sheet, the sounds are ordered in descending order of average scores across the control group (from left to right colums) and the subjects are also ordered in descending order of performance (from top to bottom rows). Only errors greater than that of the average control group have been shown.

The file exelnscat.m was used to generate this.

5. Barplots- Barplots giving the the sorted errors for each subject is given in barplots.pdf. For each subject, there are four bar plots for descending order of the error- one each for consonants in initial position, consonants in final position, vowels in initial position and vowels in final position. If there are no presentations, there is no bar displayed. For all others, the rounded percentage error is displayed above the bar, and the total number of presentations for the sound are displayed above that.

The file bar4.m was used to generate this.