Interest in Films as Measured by Subjective & Behavioral Ratings and Topographic EEG
by David Alan Kaiser
University of California, Los Angeles, 1994
Subjective interest in short films was investigated in 20 right-handed subjects (10 male, 10 female) using a Likert-type rating, a continuous rating task, and topographic EEG. Subjects watched and rated interest in approximately
15 short films. Subjects watched half of these films a second time during which they rated interest by freely moving a joystick. Baseline conditions were replicated three times. EEG was recorded from 19 cortical sites during all
conditions. Four spectral parameters of alpha activity (8-12 Hz) were calculated for two baseline conditions, a
motor control task, a film viewing task, and a film rating task. These parameters indexed amplitude, variability, trend, and trend instability. The value of these spectral parameters was discussed.
Differences in nonspecific arousal were observed between baseline and control conditions and both film tasks. Functional activation was observed in response to specific process demands in central (motor), fronto-temporal (monitoring),
parietal (narrative integration), and anterior temporal (memory) cortex. Film viewing activated right hemisphere preferentially compared to the eyes closed baseline condition. Gender differences in laterality were found in fronto-temporal
cortex in the film viewing task. Alpha activity in parietal cortex, especially the right parietal
lobe, interacted with subjective interest. Depth of narrative integration was related to left and mid-parietal activation patterns.
Behavioral interest ratings showed that interest was related to arousal modulation, as predicted by the arousal jag model proposed by Berlyne (1971). A measure of between-subject variance, called interpersonal synchrony, was developed
to identify arousal "jags" and assess subjective interest in films. The amount of interpersonal synchrony was related to subjective interest across entire films and for short intervals. Applications of these results in media research
were discussed. In addition, methodological parameters in quantitative topographic EEG were
reviewed.
TABLE OF CONTENTS
Chapters
- INTRODUCTION
- METHODOLOGICAL ISSUES
- SPECTRAL CORRELATES OF AROUSAL AND ACTIVATION
- QEEG CORRELATES OF FILM PRESENTATIONS
- QEEG CORRELATES OF SUBJECTIVE INTEREST
- ATTENTIONAL SYNCHRONY
- CONCLUSIONS
- REFERENCES
INTRODUCTION
Narrative
The narrative form is found in nearly every one of the
more than 3000 cultures that inhabit Earth. Beginning with an
oral tradition and the earliest written epics of Mesopotamia to
present-day cinema, narrative has served important psychological
and social functions. The narrative form has been wielded to
advance morality (Plato), redress injustice (Stowe), promote
tolerance (Jesus), explore the human condition (Dosteovsky), and
communicate the uncanny (Rod Serling, Borges). Feature films
(cinematic narratives) in particular immerse audience members to
such an extent that this media form has permanently transformed
the imaginations of recent generations. Evaluation of scientific
and technological progress, for example, is instinctively
compared to the ideas and images presented in "2001: A Space
Odyssey", a film now 25 years old. Steven Spielberg's film "Jaws"
ignited a fear of the ocean which has become permanent in many
individuals. Because of this persuasive power, the majority of
media research has concentrated on social, behavioral, or
cognitive consequences of viewing television and film without
considering the act of viewing itself (Blaukopf, 1990). The
physiological and psychophysiological responses, particularly
perception and integration of cinematic content, has received
little attention from scientists.
In contrast to print, television and film are generally
viewed as passive media: information is acquired (supposedly)
with little or no active participation by the viewer (Chaffee &
Schleuder, 1986). Blaukopf (1990) compares the cognitive state
created from watching television ("television stare") to the
limited conscious state of hypnosis. Yet watching a film requires
continuously and rapid integration of complex information in two
modalities. Despite evidence to the contrary (Walker, 1980;
Krugman, 1971), cinematic media has been essentially stigmatized
in terms of cognitive and attentional requirements (Greenfield &
Roos, 1988).
Behavioral and subjective measures have been used to
access attention, processing, and general responses to media with
intriguing results. Nonlinear film montages result in better
comprehension than chronological sequences (Cowen, 1988).
Conflicting information presented sequentially in film does not
demonstrate a primacy effect whereas a primacy effect is reported
for printed texts (Cowen, 1984). In addition, watching erotic
films elicits aggressive tendencies to a greater extent than
viewing aggressive films (Zillman, 1982). Consumption of
cinematic media involves a number of apparent contradictions,
including the effects on arousal. Many individuals watch
television to unwind and relax, yet just as many attend films for
excitement and emotion.
The application of psychophysiological techniques is a
new, promising technique in media research which can be used to
investigate not only nonspecific arousal but functional
correlates of media involvement. Krugman (1971) was the first to
use EEG in a media context, but only a handful of researchers
have followed up this line of research (e.g., Rothschild & Hyun,
1990; Reeves, Thorson, Rothschild, McDonald, Hirsch, & Goldstein,
1985).
Methodological issues
Numerous obstacles remain in using topographic EEG to
investigate media effects. In addition to higher costs and a
shortage of media scientists with sufficient training in
quantitative topographic EEG, the lack of methodological
standards has undermined this approach. The sensitivity and
reliability of EEG has been overlooked for cheaper, less reliable
subjective and behavioral measures. But recently more and more
researchers in various disciplines outside of neuropsychology
have embraced EEG methods as advances in technology continue to
reduce the cost of acquiring and analyzing quantitative
topographic EEG. Unfortunately, proficiency, knowledge, and
expertise in topographic EEG are not so easily acquired. Chapter
2 will examine methodological issues in continuous EEG relevant
to the study of attention and higher cognitive functions.
Handedness is also an important factor to consider in
topographic EEG research. Left handedness is associated with
different patterns of laterality than right handedness (Galin,
Ornstein, Herron, & Johnstone, 1982; Provins & Cunliffe, 1972).
Accordingly, subjects will be screened by writing samples and a
modified Edinburgh Handedness Inventory (Oldfield, 1971) and EEG
will be only be acquired from strong right-handed individuals.
Spectral parameters of activation
Measures of dispersion such as variance can be as
predictive as measures of central tendency (Cacioppo & Dorfman,
1987). Although some studies include standard deviations of alpha
activity in data tables, statistical analysis of this spectral
parameter is rarely, if ever, performed. Mean power or log power,
a single aspect of the psychophysiological state, is the one and
only measure analyzed by a majority of EEG scientists. Averaging
across tasks and conditions, a common practice in the field,
reduces the sensitivity of this measure even further. In chapter
3, two measures of dispersion of alpha activity as well as the
linear trend across conditions will be evaluated.
Also, the concept of "macrostate" underlies the
continuous EEG paradigm and most quantitative EEG research. ERP
and ERD research require hundreds of brief iterations of stimulus
presentations and recordings to ensure reliability of the EEG
signal. According to the macrostate assumption, tasks lasting a
number of minutes are associated with a distinct
electrophysiological state that can be reproduced reliability and
also reflects meaningful psychological variables. This assumption
is put to the test by each experiment presented in this work.
Processing of films
Comprehension of films requires that an individual
integrate information from two modalities simultaneously, process
plots, characters, and settings, discern thematic aspects and
fictive boundaries, as well as perform a myriad of very rapid
visual and auditory operations. Integration of such disparate
forms of information into a meaningful whole will obviously
require activation of multiple cortical functions. Chapter 4 will
investigate spectral correlates of film presentations compared to
a visual control condition as well as correlates of content
recognition.
Gender
Also in chapter 4, gender is examined as a factor
itself. Although the evidence is not conclusive, anatomical,
physiological, and behavioral findings point to functional
variation between genders, especially in cerebral asymmetry.
Males exhibit stricter segregation of most functions (McGlone,
1978; Sundet, 1986; Inglis & Lawson, 1982) whereas females
demonstrate bilateral functional representation (Turkheimer &
Farace, 1992). Although differences in alpha power and peak
frequency have also been reported (Etevenon, 1986; Shepherd,
1982; Deakin & Exley, 1979), most gender research using
topographic EEG focuses on functional laterality (McGlone, 1980).
Gender effects, however, may reflect strategic or social
differences more than physiological or anatomical differences
(Kinsbourne, 1980; Faber-Clark & Moore, 1983; Inglis & Lawson,
1982). Gender differences vary between researchers and interact
with variables such as task difficulty (Rippon, 1990; Moore &
Haynes, 1980; Shepherd, 1982). Whether gender effects in EEG
reflect absolute functional differences or task-related
divergence, further study of this variable is warranted.
Subjective Interest
The investigation of "interestingness" suffers from a
definition problem that in not uncommon in cognitive psychology
(i.e., consciousness research). Interest in a stimulus or task is
associated with the numerous properties: exciting, challenging,
and attention-holding (Wicker, Brown, & Paredes, 1990), novel,
active, concrete and personally-involving (Garner, 1992), as
vividness, narrativity, and importance (Shimoda, 1993). Some
scientists view the study of interest as a special case of
motivation (content-oriented), or embedded in larger theories of
learning and information acquisition (Iran-Nejad & Cecil, 1992),
or as property associated with both cognition and affect (Krapp,
Hidi, & Renninger, 1992).
According to Berlyne's psychobiological model of
aesthetic preference, individuals seek a momentary elevation in
arousal in order to experience the ensuing drop in arousal, which
is pleasurable (Berlyne, 1971). A stimulus is interesting if it
can induce a temporary rise in arousal, followed by the hedonic
arousal reduction. Spectral correlates of these "arousal jags"
will be searched for in ongoing EEG signal.
In chapter 5, subjective interest in film previews will
be assessed by a subjective measure, a behavioral measure, and a
psychophysiological measure. Subjective interest in a 2-min film
will be determined by self-reports using Likert-type scales. A
continuous joystick task will be used to identify interest
changes within each film. EEG derivations will be compared to
both measures.
In chapter 6, a new measure of engagement called
attentional synchrony (Async) is described. This measured
reflects processing convergence or divergence between
individuals. Unfortunately, this measure may not be applicable to
many task. It requires a large groups of subject and, more
problematic, an exact timing of stimulus presentations.
Nevertheless, this technique suggests a new approach in the study
of higher cognitive functions with quantitative topographic EEG.
|
