Toward Fully Automated PersonIndependent Detection of Mind Wandering

Toward Fully Automated Person-Independent Detection of Mind Wandering Robert Bixler & Sidney D’Mello rbixler@nd. edu University of Notre Dame July 10, 2013

mind wandering § indicates waning attention § occurs frequently § § 20 -40% of the time decreases performance § § comprehension memory

solutions § proactive § mindfulness training § § tailoring learning environment § § Mrazek (2013) Kopp, Bixler, D’Mello (2014) reactive § mind wandering detection

our goal is to detect mind wandering

related work – attention § Attention and Selection in Online Choice Tasks § § Multi-mode Saliency Dynamics Model for Analyzing Gaze and Attention § § Navalpakkam et al. (2012) Yonetani, Kawashima, and Matsuyama (2012) distinct from mind wandering

mind wandering detection § neural activity § physiology § acoustic/prosodic § eye movements

neural activity Experience Sampling During f. MRI Reveals Default Network and Executive System Contributions to Mind Wandering § Christoff et al. (2009)

physiology Automated Physiological-Based Detection of Mind Wandering during Learning § Blanchard, Bixler, D’Mello (2014)

acoustic-prosodic In the Zone: Towards Detecting Student Zoning Out Using Supervised Machine Learning § Drummond and Litman (2010)

eye movements mindless reading mindful reading

research questions 1. can mind wandering be detected from eye gaze data? 2. which features are most useful for detecting mind wandering?

data collection § 4 texts on research methods § § self-paced page-by-page 30 -40 minutes difficulty and value auditory probes § § tobii tx 300 9 per text inserted psuedorandomly (4 -12 s) type of report yes end-of-page 209 within-page 1278 no total 651 2839 860 4117

data analysis 1. compute fixations § OGAMA (Open Gaze and Mouse Analyzer) (Voßkühler et al. 2008) 2. compute features 3. build supervised machine learning models

features § global § local § context

global features § eye movements § § § fixation duration saccade length fixation dispersion reading depth fixation/saccade ratio

local features § reading patterns § § § word length hypernym depth number of synonyms frequency fixation type § § § regression first pass single gaze no word

context features § positional timing § § previous page times § § § since session start since text start since page start average previous page to average ratio task § § difficulty value

supervised machine learning § parameters § § § window size (4, 8, or 12) minimum number of fixations (5, 1/s, 2/s, or 3/s) outlier treatment (trimmed, winsorized, none) feature type (global, local, context, combined) downsampling feature selection § classifiers (20 standard from weka) § leave-several-subjects-out cross validation (66: 34 split)

1. can mind wandering be detected using eye gaze data? best model kappas 0, 3 kappa 0, 25 0, 2 0, 15 0, 1 0, 05 0 End-of-page Within-page report type

1. can mind wandering be detected using eye gaze data? 75 70 accuracy % 65 60 Accuracy 55 Expected Accuracy 50 45 40 End-of-page Within-page

1. can mind wandering be detected using eye gaze data? confusion matrices end-of-page actual classified response e within-page prior actual classified response e prior yes no yes. 54. 46. 23 yes no yes. 61. 39. 36 no. 23. 77 no. 42. 58. 64

2. which features are most useful for detecting mind wandering? kappa 0, 3 average kappa values across feature types Global 0, 2 Local 0, 1 Context 0 End-of-page Within-page report type

2. which features are most useful for detecting mind wandering? rank end-of-page within-page 1 previous value saccade length max 2 previous difficulty saccade length median 3 difficulty fixation duration ratio 4 value saccade length range 5 saccade length max saccade length mean 6 saccade length range saccade length skew 7 page number fixation duration median 8 saccade length sd fixation duration mean 9 saccade length mean saccade duration mean 10 saccade length skew saccade duration min

summary § mind wandering detection is possible § § § kappas of. 28 to. 17 end-of-page models performed better global features were best § exception: context features highest ranked for endof-page

enhanced feature set § global § § pupil diameter blink frequency saccade angle local § § cross-line saccades end-of-clause fixations

enhanced feature set 0, 3 kappa 0, 25 Original Enhanced 0, 2 0, 15 0, 1 End-of-page Within-page

predictive validity mw rate end-of-page predicted actual (model) post transfer knowledge learning -. 556 -. 248 -. 415 -. 266 actual (all data) -. 239 -. 207 within-page predicted actual (model) actual (all data) -. 496 -. 095 -. 255 -. 431 -. 090 -. 207

self-caught mind wandering kappa self-caught vs. probe caught 0, 35 0, 3 0, 25 0, 2 0, 15 0, 1 0, 05 0 End-of-page Within-page report type Self-Caught

what does mind wandering look like? § saccades § § slower shorter § more frequent blinks § larger pupil diameters

limitations § eye tracker cost § population validity § self-report § classification accuracy

future work § multiple modalities § different types of mind wandering § mind wandering intervention

acknowledgements § § § Blair Lehman Art Graesser Jennifer Neale Nigel Bosch Caitlin Mills

questions ?