Conference PaperPDF Available

The efficacy of psychophysiology for realising affective computing

Authors:
Louise Venables
Louise Venables
Louise Venables
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Jen Allanson at TupleSpace
Jen Allanson
  • TupleSpace
Stephen Fairclough at Liverpool John Moores University
Stephen Fairclough
Download full-text PDF
Read full-text
Download citation

Abstract

This demonstration reports results from the EU-funded project Ambient Agoras, investigating future applications of ubiquitous and ambient computing in workspaces. Instead of presenting underlying system technologies or evaluation findings, this demonstration ...
Content uploaded by Stephen Fairclough
Author content
All content in this area was uploaded by Stephen Fairclough on Feb 09, 2014
Content may be subject to copyright.
The Efficacy of Psychophysiology for Realising
Affective Computing
Louise Venables
Liverpool John Moores University,
Liverpool, UK
L.Venables@livjm.ac.uk
Jennifer Allanson
Liverpool John Moores University,
Liverpool, UK
J.Allanson@livjm.ac.uk
Stephen Fairclough
Liverpool John Moores University,
Liverpool, UK
S.Fairclough@liv.ac.uk
Categories & Subject Descriptors: H.1.2 [User/machine
systems]: Human Factors
General Terms: Human Factors; Experimentation.
Keywords: Affective computing; Affect recognition; Psy-
chophysiology.
INTRODUCTION
Affective computing research aims to consider human emo-
tions in the design of new computer systems. One mecha-
nism for realising this involves detectable (real-time) psy-
chophysiology as a measure of human affect [1].
Existing affective computing research has so far focused on
a small number of psychophysiological variables (PPVs):
skin conductivity (GSR), peripheral blood flow measure-
ment (BVP), and muscle tension (EMG). However, many
other PPVs are sensitive to changes in operator state [2],
(e.g. workload, stress, engagement, and fatigue), and so
may also be sensitive to aspects of human affect. These
include measures of brain activity (EEG), eye blink activity
(EOG), respiration, and cardiovascular activity (ECG). The
current study will therefore examine a range of PPVs for an
Affective Computing system in terms of their ability to
identify and predict changes in a user’s subjective state.
METHODS
To encourage negative user affect (i.e. disengagement, dis-
tress and fatigue), thirty participants performed the (3-task,
high demand) Multi Attribute Task Battery [3] for a total of
100mins. Subjective state was assessed using the Dundee
Stress State Questionnaire [4], which examines mood, mo-
tivation, thinking style and thinking content, and the Task-
Induced Fatigue Scale [5]. The psychophysiological meas-
ures recorded while the task was performed include EEG,
ECG, EOG, GSR, and respiration rate.
RESULTS
The subjective reports confirm that negative affect in-
creased over time. This was accompanied by various psy-
chophysiological changes, e.g. a decrease in heart rate and
skin resistance, and longer eye blink duration. This suggests
that PPVs other than BVP and EMG can be used to detect
changes in user state. To examine the best predictors of
subjective state, multiple regression analyses were per-
formed on the data. The results highlight a number of PPVs
as predictive of specific subjective components at particular
stages throughout the task. For example, GSR and EEG
activity seem to be strong predictors of subjective task-
engagement during the last task session.
CONCLUSIONS
This study identifies several possible candidates for an af-
fective computing system although more research is needed
to validate their sensitivity. The establishment of a set of
useful and useable physiological parameters that can be
utilised to adjust system interface and/ or functionality, will
also impact on the research of Usability Engineers, and
those concerned with the development of adaptive interac-
tive computer systems (i.e. videogames).
ACKNOWLEDGEMENTS
This project is funded by the ESPRC (R34505).
REFERENCES
1. Hudlicka, E. (2003) To feel or not to feel: The role of
affect in human-computer interaction. Int. J. Human-
Computer Studies, 59, 1-31.
2. Andreassi, J.L. (2000) Psychophysiology: Human Be-
havior and Physiological Response 4th ed. Hillsdale,
New Jersey: Earlbaum
3. Comstock, J. R. J., & Arnegard, R. J. (1992). The
Multi-Attribute Test Battery for human operator work-
load and strategic behaviour research (No. 104174):
NASA
4. Matthews, G. Joyner, L., Gilliland, K., Campbell, S.
Huggins, J., & Falconer, S. (1999). Validation of a
comprehensive stress state questionnaire: Towards a
state “big three”? In I. Mervielde, I.J. Deary, F. De-
Fruyt, & F. Ostendorf (Eds.), Personality psychology
in Europe 7, 335-350.
5. Matthews, G., & Desmond, P. A. (1998). Personality
and multiple dimensions of task-induced fatigue: A
study of simulated driving. Personality & Individual
Differences, 25, 443-458.
Copyright is held by the author/owner(s).
CHI 2004, April 24–29, 2004, Vienna, Austria.
ACM 1-58113-703-6/04/0004.
CHI 2004 ׀ Late Breaking Results Poster 24-29 April ׀ Vienna, Austria
1543
Using Pupil Size Variation during Visual Emotional Stimulation in Measuring Affective States of Non Communicative Individuals
Conference Paper
  • Jul 2013
  • Lect Notes Comput Sci
In this paper we describe an exploratory experiment conducted in the early stages of the design and development of a screening program for noncommunicative individuals. The system incorporates eye-tracking and biosensor technologies for objectively measuring affective states by combining measures of ElectroDermal Activity (EDA), temperature and eye gaze metrics of pupil dilation. The objective is to provide individuals with cognitive disabilities who experience difficulties in expressing their feelings with an alternate form of communicating their emotional states in clinical screening and occupational therapy sessions. The interactive computer-based screening program is comprised of stimuli-sets designed to elicit emotional response from viewers with a trend analysis component designed with an interface for health care providers. The focus of this paper is the feasibility of including pupil dilation as a measure reflecting affective states of individuals in the overall emotional intelligence screening system.
View
Psychophysiology: Human Behavior and Physiological Response (4th ed.)
Article
Full-text available
  • Jul 2001
Reviews the book, Psychophysiology: Human behavior and physiological response (4th ed.) by John L. Andreassi (see record 2000-07165-000). This is the fourth edition of a well-known and highly acclaimed book. Although the author modestly states, "The plan of this book is to provide students with elementary information...," the book is almost an encyclopaedic account of this extremely extensive field. Given this comprehensiveness, the book represents an excellent introduction and overview to the whole field of psychophysiology. There are two introductory chapters, the first of which provides a general introduction to the field of psychophysiology, starting very conventionally with definitions and followed by a short historical account. The following chapter provides information on the basic anatomy and physiology of the nervous system, as well as the principles behind the main techniques used. The remaining 17 chapters are grouped into three main conceptual areas: (1) measures of central nervous system activity, (2) measures of the activity of behaviourally relevant peripheral systems, and (3) applications and methodological issues including an appendix dealing with the extremely important issue of laboratory safety. Andreassi has written a comprehensive guide on research methods in psychophysiology, clearly written and very suitable for the group for whom the book is primarily intended: students of behavioural science. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
View
Personality and multiple dimensions of task-induced fatigue: A study of simulated driving
Article
Full-text available
  • Sep 1998
  • PERS INDIV DIFFER
This article reports the development of a multidimensional measure of subjective fatigue states, and its associations with personality in an experimentally-controlled context. In a study of simulated driving, 256 subjects completed a new 24-item fatigue scale as well as other subjective state measures, before and after performing a fatiguing drive. An item factor analysis identified four correlated dimensions: visual fatigue, muscular fatigue, boredom and malaise. The scales were sensitive to increased fatigue following the fatiguing drive, and showed a high degree of internal consistency. The fatigue scales correlated substantially with general state measures, such as mood and motivation. A factor analysis of fatigue and other state scales identified second-order factors of task disengagement (including boredom), physical fatigue (including the other three fatigue scales), and a distress factor. The fatigue scale was also correlated with the EPQ-R and with a measure of traits related specifically to driving, the Driving Behaviour Inventory (DBI), which includes a Fatigue Proneness scale. Bivariate and multivariate analyses showed that Fatigue Proneness was the strongest single predictor of task-induced fatigue symptoms, as predicted from an interactionist analysis of relationships between traits and states. However, the relationship between traits and states associated with fatigue was complex, and other EPQ-R and DBI traits, including neuroticism, were independently associated with fatigue.
View
To feel or not to feel: The role of affect in human-computer interaction
Article
Full-text available
  • Jul 2003
The past decade has witnessed an unprecedented growth in user interface and human–computer interaction (HCI) technologies and methods. The synergy of technological and methodological progress on the one hand, and changing user expectations on the other, are contributing to a redefinition of the requirements for effective and desirable human–computer interaction. A key component of these emerging requirements, and of effective HCI in general, is the ability of these emerging systems to address user affect. The objective of this special issue is to provide an introduction to the emerging research area of affective HCI, some of the available methods and techniques, and representative systems and applications.
View
View
The multi-attribute task battery for human operator workload and strategic behavior research
Article
  • Feb 1992
The Multi-Attribute Task (MAT) Battery provides a benchmark set of tasks for use in a wide range of lab studies of operator performance and workload. The battery incorporates tasks analogous to activities that aircraft crewmembers perform in flight, while providing a high degree of experimenter control, performance data on each subtask, and freedom to use nonpilot test subjects. Features not found in existing computer based tasks include an auditory communication task (to simulate Air Traffic Control communication), a resource management task permitting many avenues or strategies of maintaining target performance, a scheduling window which gives the operator information about future task demands, and the option of manual or automated control of tasks. Performance data are generated for each subtask. In addition, the task battery may be paused and onscreen workload rating scales presented to the subject. The MAT Battery requires a desktop computer with color graphics. The communication task requires a serial link to a second desktop computer with a voice synthesizer or digitizer card.
View
MAT - MULTI-ATTRIBUTE TASK BATTERY FOR HUMAN OPERATOR WORKLOAD AND STRATEGIC BEHAVIOR RESEARCH
Article
  • Feb 1994
MAT, a Multi-Attribute Task battery, gives the researcher the capability of performing multi-task workload and performance experiments. The battery provides a benchmark set of tasks for use in a wide range of laboratory studies of operator performance and workload. MAT incorporates tasks analogous to activities that aircraft crew members perform in flight, while providing a high degree of experiment control, performance data on each subtask, and freedom to use non-pilot test subjects. The MAT battery primary display is composed of four separate task windows which are as follows: a monitoring task window which includes gauges and warning lights, a tracking task window for the demands of manual control, a communication task window to simulate air traffic control communications, and a resource management task window which permits maintaining target levels on a fuel management task. In addition, a scheduling task window gives the researcher information about future task demands. The battery also provides the option of manual or automated control of tasks. The task generates performance data for each subtask. The task battery may be paused and onscreen workload rating scales presented to the subject. The MAT battery was designed to use a serially linked second computer to generate the voice messages for the Communications task. The MATREMX program and support files, which are included in the MAT package, were designed to work with the Heath Voice Card (Model HV-2000, available through the Heath Company, Benton Harbor, Michigan 49022); however, the MATREMX program and support files may easily be modified to work with other voice synthesizer or digitizer cards. The MAT battery task computer may also be used independent of the voice computer if no computer synthesized voice messages are desired or if some other method of presenting auditory messages is devised. MAT is written in QuickBasic and assembly language for IBM PC series and compatible computers running MS-DOS. The code in MAT is written for Microsoft QuickBasic 4.5 and Microsoft Macro Assembler 5.1. This package requires a joystick and EGA or VGA color graphics. An 80286, 386, or 486 processor machine is highly recommended. The standard distribution medium for MAT is a 5.25 inch 360K MS-DOS format diskette. The files are compressed using the PKZIP file compression utility. PKUNZIP is included on the distribution diskette. MAT was developed in 1992. IBM PC is a registered trademark of International Business Machines. MS-DOS, Microsoft QuickBasic, and Microsoft Macro Assembler are registered trademarks of Microsoft Corporation. PKZIP and PKUNZIP are registered trademarks of PKWare, Inc.
View
Validation of a comprehensive stress state questionnaire: Towards a state "big three
  • Jan 1999
  • 335-350
  • G Matthews
  • L Joyner
  • K Gilliland
  • S Campbell
  • J Huggins
  • S Falconer
Matthews, G. Joyner, L., Gilliland, K., Campbell, S. Huggins, J., & Falconer, S. (1999). Validation of a comprehensive stress state questionnaire: Towards a state "big three"? In I. Mervielde, I.J. Deary, F. De-Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe 7, 335-350.