Schedule/Lab Book

 

Week 1 (April 1-6)

-Background reading (McAdams, Krumhansl)

 

Week 2 (April 7-13)

-Possible variable dimensions for behavioral experiment

 

Week 3 (April 14-20)

-General experiment design for behavioral experiment

 

Week 4 (April 21-27)

-Created stimuli for behavioral experiment

 

Week 5 (April 28-May 4)

-Problems with behavioral experiment; trouble scaling within and between dimensions

 

-Considering MDS experiment on folk songs

 

-Discussion regarding general issues with classifying differences in EEG that can always be easily obtained from behavioral experiments

 

-Reached decision to run EEG experiments first in order to ensure that the selected stimuli are clearly classifiable via brain waves.  The behavioral experiments will take place upon completion of the EEG/classification experiments.

 

Week 6 (May 5-11)

May 5:         Combining 220C with research time for EEG experiment

 

May 6:         Meeting with Hiroko for MATLAB GUI for eventual behavioral experiment

 

May 7:         Allowable stimuli combinations matrix for behavioral experiment

 

                   More MATLAB GUI help

 

                   Suggested reading:

                             F.J. McGuigan – Experimental Psychology: Methods of Research

                             Brian C.J. Moore – Psychology of Hearing

                             Douglas Mook – Classic Experiments in Psychology

 

May 8:         Designed 2 sets of stimuli for EEG experiments

 

                   Created .wav files for intervals experiment

 

                   Determined general experiment parameters

 

May 10:       Ran subject 1 on Intervals1              experiment notes

May 11:       Ran subject 2 on Intervals1   

 

Week 7 (May 12-18)

May 12:       Began running data on Intervals1 S1

 

Some analysis notes

 

May 13:       More analysis notes

 

May 14:       Classification rates for Intervals1 S1

 

                   Trying to address lack of loudness contrast in clarinet stimuli: will record louder

                   loud stimuli (at +3dB) and softer soft (at -15dB vs. current -12dB) at various computer             

loudness levels and scale based on playback on stim machine in lab.

 

                   New .wav files for intervals1.1 – creating new .exp and .sce tiles in Presentation, and new

                   data file on krink.

 

Possible dimensions for melodic stimuli: timbre, register, forward vs inversion, rhythmic pattern

 

                   Started analysis on Intervals1 S2

 

May 16:       Ran subject 3 on Intervals1.1           experiment notes

         

                   Removing timing as dimension for melody experiment – replacing with loudness

                   for congruence with intervals experiment.

 

                   Classification rates for Intervals1 S2

 

May 18:       Ran subject 4 on Intervals1.1

 

                   Revising melodic stimuli for 1)tuning; 2)coherence with interval stims.  Stimuli will share

                   the following dimensional settings across experiments:

                                                Int/Mel       Register      Timbre        Loudness

Stim 1:        P5/orig.      high            pluck          loud

                             Stim 2:        m2/inv.      low              clarinet       soft

                             Stim 3:        P5/orig.      low              pluck          soft

                             Stim 4:        m2/inv       high            clarinet       loud

 

                   General experiment parameters for melody experiment.

 

Week 8 (May 19-25)

May 19:       More experiment parameters for melody experiment

         

                   Final .wav files for melody experiment (based on conditions in lab):

                             1_vHigh_plus3.wav

                             LMRT_v1_minus6.wav

                             LR_v1_minus12.wav

                             MT.wav

 

                   Classification rates for Intervals1.1 S3

 

                   Possible next round of stimuli for intervals experiment [not used]

                             Take Stim 1 and vary by one parameter at a time. Would give:

                             Stim 1:        P5               high            pluck          loud

                             Stim 2:        m2              high            pluck          loud

                             Stim 3:        P5               low              pluck          loud

                             Stim 4:        P5               high            clarinet       loud

                             Stim 5:        P5               high            pluck          soft

 

                   Ran subject 1 on Melody1.1

 

May 20:       Classification rates for Intervals1.1 S4

 

                   Running subject 2 on Melody1.1 **data not used, see experiment notes

 

                   Started data analysis on Melody1.1 S1

 

                   Need to create visual mappings of pairs with distances for analysis of significant stimuli

                   features

 

May 21:       Classification rates on Melody1.1 S1

 

                   Ran new subject 2 on Melody1.1

 

                   Started data analysis on Melody1.1 S2

         

                   Midi pitches of stimuli (for notation):

                             Melody high: 84-96

                             Melody low:  60-72

                             Intervals high P5:  69, 76

                             Intervals low P5:  45, 52

                             Intervals high m2:  69, 70

                             Intervals low m2: 45, 46

 

                   Ran subject 3 on Melody1.1

 

                   Compiled all pair rates for intervals experiment

 

May 22:       Classification rates for Melody1.1 S2

 

Started data analysis on Melody1.1 S3     

 

May 23:       Classification rates on Melody1.1 S3

 

May 24:       Ran subject 4 on Melody1.1

 

                   Started data analysis on Melody1.1 S4

 

                   Meeting about general poster layout – notes

 

May 25:       Classification rates on Melody1.1 S4

 

                   Starting individual ICA source classification on Melody1.1 S4

 

                   More poster discussion

 

Week 9 (May 26-June 1)

May 26-28:           Got back ICA source rates for Melody1.1 S4

         

                             Ran ICA source classification for Intervals1 S2

 

                             Ran channel-by-channel classification for Intervals1 S2

                                      -not going to use this data for now because it includes eye artifacts

 

                             Finished poster; getting printed

 

May 29:                Starting ICA source classification on Intervals1 S1

 

May 30:                Got back ICA source rates for Intervals1 S1

         

                             Started ICA source classification on Intervals1.1 S4

 

May 31:                Got ICA maps for subjects with completed ICA classification done so far

 

                             Got back ICA source rates for Intervals1.1 S4

         

                             Started ICA source classification for Intervals1.1 S3 – can’t open .mat file

                             Other unopenable .mat files: Melody1.1 S1 and S2

                             Instead starting ICA source classification for Melody1.1 S3

 

                             Created new Presentation scripts for evaluating stimuli in lab (stim # displays

                             on screen while the .wav file plays)

 

June 1:                 Got back ICA source rates for Melody1.1 S3

 

                             Can now open .mat file for Melody1.1 S2 – starting ICA source classification

 

Week 10 (June 2-8)

June 2:       Got back ICA source rates for Melody1.1 S2

 

.mat file for Intervals1.1 S3 due to removal of F8 electrode (channel 6), which changed the matrix dimensions. Channel 28 (POz) missing as well for some reason.

          -resetting ICA for subject – need to remake ICA at some point and reclassify

         

                   Discussion regarding ICA and data

 

                   Getting confusion matrices for all subjects

 

                   Started ICA source classification on Melody1.1 S1

 

June 4:       lost data for Melody1.1 S1 (MATLAB closed) – starting again

 

June 5:       Got back ICA source rates for Melody1.1 S1

 

                   Re-running pair classification for Melody1.1 S4 to confirm that stimulus numbering

                   was correct (that 2 vs 1 actually meant 22 vs 21)

 

                   Meeting with Pat re. data – see Miller and Nicely (1955)

 

June 6:       Got back new pair classification for Melody1.1 S4

 

                   Remaking ICA for Intervals1.1 S3

 

June 7:       Started 4-class classification (2nd run) on Intervals1.1 S3

 

June 8:       Started ICA source classification on Intervals1.1 S3

 

Home