Under Pressure Sensing Stress of Computer Users Javier Hernandez - - PowerPoint PPT Presentation

▶
under pressure
SMART_READER_LITE
LIVE PREVIEW

Under Pressure Sensing Stress of Computer Users Javier Hernandez - - PowerPoint PPT Presentation

Sep 25, 2022 437 likes •744 views

Under Pressure Sensing Stress of Computer Users Javier Hernandez javierhr@mit.edu Pablo Paredes pablo.paredes@berkeley.edu Asta Roseway astar@microsoft.com Mary Czerwinski marycz@microsoft.com Headaches Fatigue Anxiety Insomnia Depression 2

slide-1
SLIDE 1

Under Pressure

Sensing Stress of Computer Users

Javier Hernandez javierhr@mit.edu Pablo Paredes pablo.paredes@berkeley.edu Asta Roseway astar@microsoft.com Mary Czerwinski marycz@microsoft.com

slide-2
SLIDE 2

Depression Anxiety Fatigue Headaches Insomnia

2

slide-3
SLIDE 3

Outline

  • Stress Measurement
  • Previous Work
  • User Study
  • Results
  • Conclusions

3

slide-4
SLIDE 4

Stress Measurement

4

Behaviors Hormones

Intrusive Costly Slow Cortisol Adrenaline …

Self-reports

Subjective Cognitive Attention Recall Problems DSI PSS …

Physiology

Instrumentation Unfamiliar Caring EMG EDA … Existing Interactions Continuous Monitoring Keyboard Mouse …

(Zimmermann et al, 2003)

slide-5
SLIDE 5

Keyboard & Mouse

Pressure-sensitive Keyboard

(Dietz et al., 2009)

Capacitive Mouse

(Villar et al., 2009 Benko et al., 2010) 5

slide-6
SLIDE 6

Keyboard & Mouse

Non-Pressure Dynamics

  • Neutral/positive VS negative (Khanna et al, 2010)

More typing mistakes and slower speeds

  • 15 emotional states (Epp, 2011)
  • Cognitive VS physical stress (Vizer et al, 2009)

Pressure Dynamics

  • 6 emotions (Lv et al, 2008)

Non-Pressure Dynamics

  • Increased speed and acceleration during

high arousal (Maehr, 2008)

  • Increased movement during stress

(Rodriges et al, 2013)

Pressure Dynamics

  • Increased pressure during stress

(Wahlstrom et al 2012, Dennerlein et al, 2003)

6

slide-7
SLIDE 7

7 7

Study Design Stress Measurement Experimental Tasks Protocol

slide-8
SLIDE 8

8

Within-subject laboratory study Three tasks:

  • Expressive Writing
  • Text Transcription
  • Mouse Clicking

Two conditions Stressed VS Relaxed

8

Study Design

Stress Measurement Experimental Tasks Protocol

slide-9
SLIDE 9

9

Physiological Stress Self-reported Stress

Affectiva QTM Electrodermal Activity (Galvanic Skin Response)

9

Study Design

Stress Measurement

Experimental Tasks Protocol

slide-10
SLIDE 10

10

Task I: Text Transcription

  • Transcribe short biographical text
  • 3 minutes
  • During the stressed condition:

1. Type fast 2. Internal competition 3. Cursor blinks faster 4. Random font style 5. Timer 6. Loud traffic noise

Study Design Stress Measurement

Experimental Tasks

Protocol

slide-11
SLIDE 11

11

Task II: Expressive Writing

  • Re-experience a relaxing or stressful

memory and write about it

  • Recommended time of 5 minutes
  • Allowed to make spelling, grammar

and sentence errors

11

Study Design Stress Measurement

Experimental Tasks

Protocol

slide-12
SLIDE 12

12

Task III: Mouse Clicking

  • After keyboard tasks
  • Based on Fitts’ law task

90 times (3 dist. x 3 widths x 10 rep.)

12

Study Design Stress Measurement

Experimental Tasks

Protocol

slide-13
SLIDE 13

13 Training TT MC EW TT MC EW EW MC TT EW MC TT TT MC EW TT MC EW EW MC TT EW MC TT EW: Expressive Writing MC: Mouse Clicking TT: Text Transcription Stressed Condition Relaxed Condition 24 participants Balanced gender

Study Design Stress Measurement Experimental Tasks

Protocol

13

slide-14
SLIDE 14

Results

slide-15
SLIDE 15

Did the tasks elicit the intended emotions?

slide-16
SLIDE 16

2 3 4 5 6 7

Expressive Writing Text Transcription Mouse Clicking Very Stressed Not Stressed at All

Valence Arousal

Very Pleasant Very Unpleasant High Energy Low Energy Expressive Writing Text Transcription Mouse Clicking Expressive Writing Text Transcription Mouse Clicking

Stress

* * * * * * * *

1 7 6 5 4 3 2 1 7 6 5 4 3 2 1 6 5 4 3 2

Stressed Condition Relaxed Condition Standard Error

Self-reports

16

*The two distributions were significantly different (Wilcoxon Rank Sum test , p<0.05)

slide-17
SLIDE 17

How is typing pressure affected by stress?

slide-18
SLIDE 18
  • 8
  • 6
  • 4
  • 2

2 4 6 8 0104 0105 0108 0103 0106 0109 0203 0204 0205 0206 0207 0208 0301 0302 0303 0304 0305 0306 0401 0402 0403 0404 0405 0406 P ti i t 70 75 80 85 90 95 0104 0105 0108 0103 0106 0109 0203 0204 0205 0206 0207 0208 0301 0302 0303 0304 0305 0306 0401 0402 0403 0404 0405 0406 P ti i t

Transcription Task

Stressed condition Relaxed condition

  • Avg. Pressure
  • Avg. Stressed – Avg. Relaxed

Higher Pressure During Stressed Higher Pressure During Relaxed

x x x

18

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

* * * * * * * * * * * * * * * * * * * * *

Participants

22 out of 24 subjects (91.67%) show increased typing pressure under the stressed condition The difference was computed from significantly different distributions (Wilcoxon Rank Sum test , p<0.05)

slide-19
SLIDE 19
  • 8
  • 6
  • 4
  • 2

2 4 6 8 0104 0105 0108 0103 0106 0109 0203 0204 0205 0206 0207 0208 0301 0302 0303 0304 0305 0306 0401 0402 0403 0404 0405 0406 P ti i t 70 75 80 85 90 95 1 4 1 5 1 8 1 3 1 6 1 9 2 3 2 4 2 5 2 6 2 7 2 8 3 1 3 2 3 3 3 4 3 5 3 6 4 1 4 2 4 3 4 4 4 5 4 6 P ti i t

Stressed condition Relaxed condition

x

23 out of 24 subjects (95.83%) show increased typing pressure under the stressed condition The difference was computed from significantly different distributions (Wilcoxon Rank Sum test , p<0.05)

Expressive Writing

  • Avg. Pressure
  • Avg. Stressed – Avg. Relaxed

Higher Pressure During Stressed Higher Pressure During Relaxed

x x x

19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

* * * * * * * * * * * * * * * * * * * *

Participants

slide-20
SLIDE 20

How is mouse capacitance affected by stress?

slide-21
SLIDE 21

Capacitive Mouse

21

13x15 Capacitive Pixels

slide-22
SLIDE 22

00:00 00:07 00:15 00:22 00:30 00:37 00:45 00:52 01:00 01:07 8 10 12 14 16 18 20 22 24 FITT_1 (1) 5 10 15 2 4 6 8 10 12 14 FITT_1 (842) 5 10 15 2 4 6 8 10 12 14 FITT_1 (1682) 5 10 15 2 4 6 8 10 12 14 FITT_1 (2523) 5 10 15 2 4 6 8 10 12 14 FITT_1 (3364) 5 10 15 2 4 6 8 10 12 14 FITT_1 (4204) 5 10 15 2 4 6 8 10 12 14 FITT_1 (5045) 5 10 15 2 4 6 8 10 12 14 FITT_1 (5886) 5 10 15 2 4 6 8 10 12 14 FITT_1 (6726) 5 10 15 2 4 6 8 10 12 14 FITT_1 (7567) 5 10 15 2 4 6 8 10 12 14 FITT_2 (1) 5 10 15 2 4 6 8 10 12 14 FITT_2 (762) 5 10 15 2 4 6 8 10 12 14 FITT_2 (1523) 5 10 15 2 4 6 8 10 12 14 FITT_2 (2284) 5 10 15 2 4 6 8 10 12 14 FITT_2 (3045) 5 10 15 2 4 6 8 10 12 14 FITT_2 (3805) 5 10 15 2 4 6 8 10 12 14 FITT_2 (4566) 5 10 15 2 4 6 8 10 12 14 FITT_2 (5327) 5 10 15 2 4 6 8 10 12 14 FITT_2 (6088) 5 10 15 2 4 6 8 10 12 14 FITT_2 (6849) 5 10 15 2 4 6 8 10 12 14

Start Task

Mouse Clicking

Average Time

Stressed condition Relaxed condition

Mouse Task

22

Relaxed Stressed

… …

End Task Average

slide-23
SLIDE 23

Mouse Clicking

5 10 15 20 25 0104 0105 0108 0103 0106 0109 0203 0204 0205 0206 0207 0208 0301 0302 0303 0304 0305 0306 0401 0402 0403 0404 0405 0406 P ti i t # Seconds

  • 4
  • 2

2 4 0104 0105 0108 0103 0106 0109 0203 0204 0205 0206 0207 0208 0301 0302 0303 0304 0305 0306 0401 0402 0403 0404 0405 0406 Mean Stress - Mean Relax

Stressed condition Relaxed condition

  • Avg. Mouse Contact
  • Avg. Stressed – Avg. Relaxed

More Mouse Contact During Stressed More Mouse Contact During Relaxed

18 out of 24 subjects (75%) show increased mouse contact under the stressed condition The difference was computed from significantly different distributions (Wilcoxon Rank Sum test , p<0.05) 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

* * * * * * * * * * * * * * * * * * * * * * * *

Participants

slide-24
SLIDE 24

How much data is required to differentiate between the stressed and relaxed conditions at any point in time?

slide-25
SLIDE 25

25

50 100 150 200 250 300 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Average # Participants with significantly more pressure/contact during the stressed condition Text Transcription Mouse Clicking Expressive Memory Observation Window (seconds) N = 500 randomly segmented windows Stabilizes after 3.37 seconds Stabilizes after 2.5 minutes

slide-26
SLIDE 26

Considerations

  • Ill-defined ground truth
  • Laboratory
  • Many types of stress
  • Direction of causality
  • Individual differences

26

slide-27
SLIDE 27

Conclusions

  • Lab study to test effectiveness of pressure-sensitive keyboard and capacitive

mouse to sense stress of computer users

  • 3 tasks (expressive writing, text transcription, and mouse clicking)

under 2 conditions (stressed and relaxed)

  • Self-reports showed the effectiveness of tasks
  • During the stressed condition:

– >79% showed more forceful typing pressure – 75% showed greater amount of mouse contact

  • Very small observation windows can be used to obtain similar results

27

slide-28
SLIDE 28

Acknowledgements

28

Michel Pahud Hrvoje Benko Paul H. Dietz Kevin Hinckley

slide-29
SLIDE 29

Under Pressure

Sensing Stress of Computer Users

Javier Hernandez javierhr@mit.edu Pablo Paredes pablo.paredes@berkeley.edu Asta Roseway astar@microsoft.com Mary Czerwinski marycz@microsoft.com