Wertheim's hypothesis on 'highway hypnosis': Empirical evidence from a study on motorway and conventional road driving

Abstract

This paper aims to study the phenomenon known as 'highway hypnosis' or 'driving without attention mode', which has been defined as a state showing sleepiness signs and attention slip resulting from driving a motor vehicle for a long period in a highly predictable environment with low event occurrence, this being the case with motorways and very familiar roads [Highway hypnosis: a theoretical analysis. In: Gale, A.G., Brown, I.D., Haslegrave, C.M., Moorhead, I., Taylor, S. (Eds.), Vision in Vehicles-III. Elsevier, North-Holland, pp. 467-472]. According to Wertheim's hypothesis on 'highway hypnosis', long-term driving on motorways and conventional roads, e.g. main roads, secondary roads--implies differences in the predictability of the movement pattern of the visual stimulation, in the eye musculature activity and in the type of feedback used in visual information processing (mostly extra-retinal on motorways and retinal and extra-retinal on conventional roads). All this ultimately leads to alertness differences between both road types. Our research is intended to provide empirical evidence from the hypothesis, based on the data recorded during the actual driving experience of a group of subjects on a motorway and a conventional road. We studied whether or not significant alertness differences were found-measured by EEG data relative to time periods of on-target eye-tracking performance--between motorway and conventional road driving. Our results partially support the hypothesis, as drowsiness proved to be higher on motorways than on conventional roads during the final driving period but not during the starting stage, when the opposite trend was noticed. This result could be explained by the fact that during the first driving periods the effects of the stimulus movement predictability had not yet become apparent, since they tend to show after a long drive.

Full text

Accident Analysis and Prevention 36 (2004) 1045–1054
Wertheim’s hypothesis on ‘highway hypnosis’: empirical evidence
from a study on motorway and conventional road driving
Gemma Pastor Cerezuela∗, Pilar Tejero, Mariano Chóliz,
Mauricio Chisvert, M. José Monteagudo
Instituto Universitario de Tráfico y Seguridad Vial (INTRAS), University of Valencia, Valencia, Spain
Received 18 March 2003; received in revised form 23 November 2003; accepted 11 February 2004
Abstract
This paper aims to study the phenomenon known as ‘highway hypnosis’ or ‘driving without attention mode’, which has been defined
as a state showing sleepiness signs and attention slip resulting from driving a motor vehicle for a long period in a highly predictable
environment with low event occurrence, this being the case with motorways and very familiar roads [Highway hypnosis: a theoretical
analysis. In: Gale, A.G., Brown, I.D., Haslegrave, C.M., Moorhead, I., Taylor, S. (Eds.), Vision in Vehicles-III. Elsevier, North-Holland,
pp. 467–472]. According to Wertheim’s hypothesis on ‘highway hypnosis’, long-term driving on motorways and conventional roads,
e.g. main roads, secondary roads—implies differences in the predictability of the movement pattern of the visual stimulation, in the eye
musculature activity and in the type of feedback used in visual information processing (mostly extra-retinal on motorways and retinal and
extra-retinal on conventional roads). All this ultimately leads to alertness differences between both road types. Our research is intended
to provide empirical evidence from the hypothesis, based on the data recorded during the actual driving experience of a group of subjects
on a motorway and a conventional road. We studied whether or not significant alertness differences were found-measured by EEG data
relative to time periods of on-target eye-tracking performance—between motorway and conventional road driving. Our results partially
support the hypothesis, as drowsiness proved to be higher on motorways than on conventional roads during the final driving period but
not during the starting stage, when the opposite trend was noticed. This result could be explained by the fact that during the first driving
periods the effects of the stimulus movement predictability had not yet become apparent, since they tend to show after a long drive.
© 2004 Elsevier Ltd. All rights reserved.
Keywords: Driving; Motorway; Conventional road; Alertness; Highway hypnosis; Inattention; Oculomotor control; Time on task
1. Introduction
‘Highway hypnosis’ (Williams, 1963) or ‘driving without
attention mode’ (Kerr, 1991) has been defined as a state
showing sleepiness signs and attention slip resulting from
driving a motor vehicle for a long period in a highly pre-
dictable environment with low event occurrence (Wertheim,
1991). This condition entails different physical, subjective,
behavioural and psychophysiological signs. At the physical
level, drivers tend to feel sleepy or drowsy even though
they remain seated in a normal position, with hands on the
steering wheel and looking forward (Brown, 1991). Their
gaze, however, has no expression and has been described as
a ‘glassy stare’ (Williams, 1963), this being one of the main
∗Corresponding author. Present address: Departamento de Psicolog´
ıa
B´
asica, Facultad de Psicolog´
ıa, (Universidad de Valencia), Av. Blasco
Ib´
añez 21, Valencia 46010, Spain. Tel.: +34-96-3864447;
fax: +34-96-3864822.
E-mail address: gemma.Pastor@uv.es (G.P. Cerezuela).
features of the state. In subjective terms, nothing seems to
disrupt the driver’s attention, but an ‘apparent conscious-
ness loss’ takes place in such a way that driving becomes
automatic (Kerr, 1991). Once attention is regained, the
driver hardly remembers anything about the way already
travelled. Regarding behaviour, the driver can continue
driving in global terms and can even feel that his strength
will not fail him, which probably adds to the difficulty in
realizing the lack of attention that conditions the driving
(Wertheim, 1991). In psychophysiological terms, this con-
dition is put down to a reduced arousal level resulting in
waning alertness, this in turn increasing the probability of
risky driving (Wertheim, 1978).
Different explanations come to support ‘highway
hypnosis’. The condition has been associated to factors such
as monotony, mental fatigue, and even the existence of a
disease (for review, see: Williams and Shor, 1970; Shor and
Thackray, 1970; Brown, 1994, 1997; Hancock and Verwey,
1997; Pastor et al., 1999). According to Wertheim (1991)
0001-4575/$ – see front matter © 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aap.2004.02.002

1046 G.P. Cerezuela et al. / Accident Analysis and Prevention 36 (2004) 1045–1054
and other authors (Kerr, 1991; Brown, 1991, 1994), some
elements in the road environment and the manner in which
they affect the operation of the oculomotor system could
actually make up the key factor in explaining the occurrence
of this inattention state.
1.1. Wertheim’s hypothesis on ‘highway hypnosis’:
empirical evidence from experimental laboratory tests
In Wertheim’s account (1978), some mental skills are re-
lated to the activity of the oculomotor system. This neurolog-
ical system causes eye movements to start. While driving a
vehicle, the main feedback signal for the oculomotor system
comes from the retinal information on the position of a visual
stimulus (Stark, 1968; Carpenter, 1977; Zeevi et al., 1979).
However, to some extent the oculomotor system also uses
extra-retinal signals as a feedback for controlling eye move-
ments (Wertheim, 1974, 1981a, 1981b). Extra-retinal signals
would probably collect the information coming from the
mental representations and internal motor programmes (for
review, see: Wertheim, 1974). Consequently, the oculomotor
system would receive two information types: retinal (exter-
nal) and extra-retinal (internal) (Heywood, 1972; Murphy
et al., 1975; Steinbach, 1976; Keller, 1977), which allows
us to distinguish two oculomotor activity components: the
attentive component, mainly ruled by retinal feedback, and
the intentive component, mainly ruled by extra-retinal feed-
back. While driving, the attentive component is generally the
predominant type of oculomotor activity, which favours the
processing of external visual information coming from the
driver’s traffic environment (bottom-up or stimulus-guided
processing). However, under some conditions the intentive
component becomes predominant. Then, the information
coming from internal mental representations and motor
programmes is used by the oculomotor system as the main
input (top–down or memory-guided processing), at the ex-
pense of visual information from the external environment.
According to Wertheim, the inattention state classed as
‘highway hypnosis’ results from the rising dependence on
the oculomotor system of extra-retinal feedback and the
waning dependence on retinal feedback. In other words,
intentive oculomotor activity prevails more and more over
attentive oculomotor activity, this resulting from the spe-
cific environmental characteristics of motorway driving and
driving on highly familiar roads, namely those in which the
motion trajectory of the visual scene on the retina is highly
predictable. While driving on these types of roads and envi-
ronments, the high predictability of the stimulus movement
produces a rising dependence on the oculomotor control
of extra-retinal feedback and, consequently, a greater pre-
dominance of the intentive oculomotor activity in such a
way that the retinal information is not sufficiently used as
feedback on the performance. Thus, driving is undertaken
on the basis of information derived from the perception of
very few external visual signals. The visual system remains
intact but, as pointed out by Kerr (1991, p. 475), ‘the brain
is inattentive to visual inputs’. In sum, a rigid cognitive
layout is established leading to the subjective notion of a
non-changing visual environment this in turn giving way to
a rigid ‘automatic driving’ style.
Wertheim (1978, 1981a) devised a procedure to allow the
attentive and intentive components of oculomotor control
to be experimentally manipulated at the laboratory, with a
view to ascertaining whether or not different psychological
functions—e.g. alertness—were affected by the oculomotor
control type. To that end, an eye-tracking task was designed
in which experimental subjects were supposed to track
down a visual target going round in circles on a screen. The
manipulated variable was the predictability of the motion
trajectory of the visual stimulus (high versus low predictabil-
ity). Wertheim assumed that, in the high predictability
condition, the eye-tracking task would be completed pre-
dominantly on the basis of extra-retinal feedback, the inten-
tive component of the oculomotor control then growing at the
expense of the attentive. Yet, for the low predictability con-
dition, the attentive component would remain predominant
in oculomotor control, and so the eye-tracking task would
be performed on the basis of retinal feedback. Moreover,
two additional conditions were set in the eye-tracking task:
on-target (‘adequate’ tracking) and off-target (‘inadequate’
tracking), with a wiew to comparing the two predictability
conditions during the on-target ocular performance only.
To find out whether the oculomotor control type affects
the alertness level, Wertheim used a psychophysiological
measurement that can be influenced by the activity of the
visual system: EEG alpha activity (8–12Hz). This activity
is usually measured on the visual centres of the brain, which
correspond to the skull areas in the occipital–parietal region.
In general terms, alpha activity is present in states of mental
relaxation, physiological inactivity, surveillance loss, and is
either absent or blocked during visual attention (Ray, 1990).
Experimental papers abound attempting to prove the rela-
tionship between EEG alpha activity and eye movements
(e.g. Bender, 1969; Mulholland and Evans, 1965; Mulhol-
land and Peper, 1971), the same applies to theoretical formu-
lations about its involvement in attention mechanisms and
cognitive-emotional processes (for review, see: Wertheim,
1974; Plotkin, 1976; Ulrich, 1990). According to Wertheim
(1974), when the attentive component of oculomotor control
predominates, the occipital alpha activity is either softened
or blocked, this being a sign of mental alertness, whereas
when the intentive component is predominant—usually in
automatic performance with a visual task—an increase is
found in the amplitude of the occipital alpha activity, which
indicates low mental alertness. In his experimental work
(Wertheim, 1978), the occipital alpha activity was found to a
greater extent in the high predictability condition (intentive
control) than in the low predictability one (attentive control),
during the on-target ocular performance.
Should this result be applicable to an actual driving envi-
ronment, then the hypothesis—in Wertheim’s view—could
be launched that prolonged driving in settings where the

G.P. Cerezuela et al. / Accident Analysis and Prevention 36 (2004) 1045–1054 1047
motion trajectory of the stimuli on the retina is highly pre-
dictable—as is the case with motorways and familiar
roads—would cause the predominant oculomotor control
type to switch to intentive mode, driving taking place on the
basis of predominantly extra-retinal feedback. Nevertheless,
in environments where the motion trajectory of the stimuli
on the retina is less predictable—this being the case with
conventional roads, e.g. main roads and secondary roads—
driving would give way to a predominantly attentive ocu-
lomotor control type in the visual information processing,
driving occurring on the basis of retinal and extra-retinal
feedback. This implies that the alertness level—evidenced
through EEG activity—would be lower on motorways and
familiar roads than on conventional roads for those time
periods in which the driver visually tracks the trajectory of
the stimulus motion (on-target ocular performance periods).
This was used as the starting hypothesis of our empirical
research.
To date, Wertheim’s laboratory experiments seem to make
up the only empirical evidence of his hypothesis on ‘high-
way hypnosis’. However, the outcome of a recently con-
ducted driving simulation exploratory study (Thiffault and
Bergeron, 2003) on the effects of environmental monotony
and predictability upon fatigue—assessed through perfor-
mance measures—shows that certain steering behaviour
measures deteriorate to a greater extent while driving in a
monotonous, repetitive, and predictable visual environment
than in variable, non-predictable visual settings.
1.2. Aims
After Wertheim’s laboratory experiments, no subsequent
research has been found in the literature providing empirical
evidence from the hypothesis on ‘highway hypnosis’. The
author himself stressed the importance of studying the hy-
pothesis in the field, since he was aware of the constraints of
his laboratory experiments when making the results general
to the driving environment. For this reason, the target of our
research was to present empirical evidence from Wertheim’s
hypothesis on ‘highway hypnosis’ based on data recorded
in the actual driving experiences under normal conditions
of a group of volunteer subjects on motorway and con-
ventional road. The experiment was intended to point to
potential significant differences between motorway and con-
ventional road driving with regard to alertness in the road
eye-tracking periods (on-target periods). Motorway driving
represents the predictability condition whereas conventional
road driving represents the non-predictability condition, of
the trajectory of the stimulus movement produced by our
own motion, as argued by Wertheim (1978) when he talked
about the possibility of studying ‘highway hypnosis’ in the
field. Alertness was measured through EEG, selecting only
those time periods in which subjects looked forward while
driving on a motorway and a conventional road (on-target
ocular performance periods). Only during such periods
subjects are assumed to adequately track the trajectory
of the movement of the road stimuli during the motion
itself.
2. Methods
2.1. Participants
The study included 14 subjects (12 men and 2 women)
who had been previously contacted by the University In-
stitute for Traffic and Road Safety (University of Valencia,
Spain). Mean age was 29,64 years, ranging from 24 to 54.
All of them were experienced drivers (they had had a driv-
ing licence for at least 7 years, had driven 60,000km min-
imum, with a weekly driving frequency of three or more
times a week in the last year). Randomly selected, half of
the subjects drove on a motorway and the other half on a
conventional road. They were all paid for their participation
in the research.
2.2. Instruments
The car driven by the subjects was an Opel Kadett 1600,
on which several recording devices had been fitted. A Pana-
sonic video-camera was fitted inside the vehicle, at the front,
focused on the driver’s face in order to record his/her eye
movements; a Panasonic 2.5in. monitor was connected to
the video-camera with a view to ensuring that the camera was
focused on the driver’s eyes at all times. Electrophysiolog-
ical recording equipment (Biopac Systems, CA, USA) was
fitted on the other front seat to record EEG activity. Spoon
electrodes made of silver and covered with silver chloride
(Ag/AgCl) were used. They were fitted using a conductor
gel (EC2 Electrode Cream). Specially sensitive in somno-
lence recording, two EEG channels were utilised to express
sleepiness through alpha (8–12 Hz) and theta (4–8 Hz) ac-
tivity increases (Åkerstedt and Gillberg, 1990; Kecklund
and Åkerstedt, 1993): a central derivation in position C3-A2
(monopolar fitting) and an occipital–parietal derivation in
position O2-P4 (bipolar fitting) (as per electrode fitting in
the 10–20 International System). In general terms, the EEG
central derivation used by the study is considered to be
most adequate for somnolence and drowsiness recording; in
fact, it is preferentially recommended for studies on sleep
phases in humans, since it is particularly sensitive to detect
the start of the sleep, when brain activity decreases con-
siderably (Rechtschaffen and Kales, 1968). As to the EEG
occipital–parietal derivation, EEG activity on the alpha
band recorded on the occipital region is deemed to be asso-
ciated with the predominance of automatic control in road
eye-tracking (Wertheim, 1981a), thus becoming a driving
automatization index. Lastly, with a view to visualising and
storing the EEG signals obtained from the driving, a portable
Pentium PC (Dell) was used and handled by the experi-
menter at the back of the car. The computer was connected
to the Biopac Systems polygraph, in such a way that all EEG

1048 G.P. Cerezuela et al. / Accident Analysis and Prevention 36 (2004) 1045–1054
activity data were automatically stored in a file generated
by a dedicated signal acquisition software (Acqnowledge
MP100WSW, version 3.2.4.) installed in the same computer.
2.3. Procedure
Data were collected from the subjects’ driving experi-
ences on a stretch of the A-7 motorway and the N-340 road
(a main road, only one lane for each direction), in the Va-
lencian Community (Spain). Data were collected under very
similar circumstances in all cases, always at the same time
of the day, with medium traffic, and comparable weather and
light conditions. Subjects were told to keep a constant speed
between 100 and 120km/h, to stay on the right lane, and
remain silent. Subjects were accompanied by a researcher
who was in charge of the laptop, seeing to it that the experi-
ment was conducted correctly. Once the session started, the
researcher would ask the subject to look at each of the visual
field areas that were to be considered later on in eye perfor-
mance codification: the front, the dashboard, the wing mir-
rors, and the scenery. In this manner, personalised records
could be kept for each driver as to the position of the eyes
when looking at each of the areas, this being used as a sam-
ple facilitating subsequent codification.
The beginning of the driving stretch was Valencia; it con-
tinued on the N-221 road up to the entrance to either the
A-7 or the N-340. At this point, half the subjects took the
motorway and the other half took the main road. The ap-
proximate distance to drive was 90km in both cases, thus
EEG activity data and eye performance video recording data
were recorded in a continuous uninterrupted way for about
45min. The accompanying researcher would enter—in the
same file where EEG data were automatically viewed and
stored—any driving event or incident (changing lanes, over-
taking, speed changes, road works, whether or not subjects
spoke, etc.) so that all the information on the events and the
time they had occurred was recorded on the computer.
2.4. Design and variables
A4×2 mixed factorial design was used including two in-
dependent variables: driving time period (intra-subject vari-
able) and road type (inter-subject variable). As far as the
former is concerned, driving time was split into four 10-min
periods as from minute 5 on the corresponding road (motor-
way or conventional road) with a view to comparing driving
conditions between motorway and conventional road in sim-
ilar time periods. The four time periods make up the condi-
tions of the driving time period variable. The conditions of
the road type variable are: motorway driving and conven-
tional road driving, which represent the predictability and
non-predictability of the trajectory of the movement of the
stimuli during the motion, respectively.
Dependent variables were EEG activity measurements—
central (C3-A2) and occipital–parietal (O2-P4) derivations—
corresponding to on-target ocular performance time periods.
For each derivation, two EEG activity measurements specif-
ically related to the start of drowsiness were analysed:
•Spectral density on the alpha frequency band (8–12Hz):
this absolute measurement indicates low alertness. EEG
alpha activity—especially in the occipital area—can be
considered an indicator of the automatization level in road
eye-tracking while driving (Wertheim, 1981a). The ac-
tivity of the occipital alpha band in the EEG was used
by Wertheim (1978) in his laboratory work on ‘highway
hypnosis’, showing that it was present in the predictability
condition to a greater extent than in the non-predictability
one during the on-target ocular performance.
•The relative energy parameter [(theta +alpha/beta]
(θ+α/β): this relative measurement points to low alert-
ness and so constitutes an overall drowsiness indicator.
It has been used by other driving analyses (De Waard
and Brookhuis, 1991; Brookhuis and De Waard, 1993),
and results from dividing the spectral density on the slow
EEG bands by the spectral density of the fast bands, as
shown by the following formula:
[spectral density on theta band(4–8 Hz)
+spectral density on alpha band(8–12 Hz)]
spectral density on beta band(12–30 Hz)
3. Analysis of data
3.1. Eye performance codification
Eye performance video recordings were viewed by au-
thors to determine those driving moments at which subjects
looked at some of the areas set. This was done by means
of an MSDOS computer programme consisting of a timer
to store the flag (for each area to which the eyes turned to)
and the flag time. Once the eye performance was so cod-
ified, time periods during which subjects looked forward
were determined for both motorway and conventional road
(road tracking periods or on-target ocular performance peri-
ods); the same applied to the time at which subjects did not
look forward (off-target ocular performance periods), since
the latter corresponded to the looks at any area other than
the front (wing mirrors, dashboard, scenery).
3.2. Selection of EEG activity fragments corresponding to
on-target ocular performance
Following the synchronisation of EEG activity data with
eye performance data, those time EEG activity fragments
(central and occipital–parietal derivations) corresponding
to on-target ocular performance were selected for both
motorway and conventional road; more particularly, only
those EEG fragments corresponding to forward look peri-
ods of at least 4.096s were chosen, since this is the length
of the EEG analysis periods (length of sampling intervals

G.P. Cerezuela et al. / Accident Analysis and Prevention 36 (2004) 1045–1054 1049
or epochs). As a criterion, no incidence should have been
detected during such periods (changing lanes, overtaking,
speed changes, etc.). The finally selected EEG fragments
were called on-target EEG.
3.3. Analysis of on-target EEG activity
EEG signals recorded at a sampling rate of 125 data per
second were filtered using a band-pass filter in order to
eliminate artefacts and select frequency bands between 4
and 30Hz (theta, alpha and beta bands). Spectral analysis
was performed by calculating the Fast Fourier Transform
(Gottman, 1990). Accumulated spectral density was ob-
tained for frequency bands theta (4–8Hz), alpha (8–12 Hz)
and beta (12–30Hz), and the relative energy parameter
(θ+α/β) was calculated. Calculations were obtained each
4.096 s (512 data). After that, indicators were logarithmi-
cally transformed in order to standardise their distribution.
Finally, taking only those EEG fragments that had been
selected (on-target EEG), the measurements obtained every
4.096s were grouped into 60-s (per minute) time periods.
To that end, all spectral density values within each estab-
lished minute were averaged out up until minute 45. In this
way, the mean values per minute of each on-target EEG in-
dicator were drawn for each subject and for each derivation
on both the motorway and the conventional road.
To carry out contrast statistical tests, data for each in-
dicator per minute were regrouped into four 10-min time
periods plus one 5-min baseline initial period, for motorway
and conventional road. Consequently, driving time periods
were as follows: period 1 or baseline (min 0–5), period 2
(min 5–15), period 3 (min 15–25), period 4 (min 25–35),
and period 5 (min 35–45). According to the Initial Value
-0,3
-0,2
-0,1
0
0,1
0,2
0,3
0,4
12345
Driving time period
Alpha (8-12 Hz) on-target
Motorway Conventional road
Scheme 1. Spectral density on alpha band for EEG on-target periods (central derivation).
Law (Wilder, 1950, 1965), the evolution in time of psy-
chophysiological signals in each subject is partly determined
by their initial level or baseline, which usually presents
important inter-individual differences. This is why we es-
tablished an initial period (period 1) as a baseline indicator
of each subject at each measurement. In order to control
baseline inter-individual variability, different procedures
can be observed (Cronbach and Furby, 1970; Johnson and
Lubin, 1972; Russell, 1990). Due to its straightforwardness,
we chose a method consisting of subtracting—from each
of the values in each period of the indicator—the value
corresponding to that measurement on the baseline (period
1), for each subject.
Following the univaried approach and using Ftests av-
eraged for intra-subject effects (Vasey and Thayer, 1987),
mixed variance analyses were conducted with the data from
the spectral density measurement on the alpha band and
those from the parameter (θ+α/β), of the on-target EEG
(central and occipital–parietal derivations). The overall ef-
fect of each independent variable (driving time period and
road type) and the overall influence of their interaction were
calculated. After that, any specific interaction of interest was
analysed, namely those corresponding to the effect of the
road type variable within each time period set. All analyses
were performed with the SPSS statistical package, version
8.0 for Windows.
4. Results
Let us remember that the mean values obtained on the
motorway and conventional road for the different measure-
ments in each time period are values relative to the baseline

1050 G.P. Cerezuela et al. / Accident Analysis and Prevention 36 (2004) 1045–1054
-0,25
-0,2
-0,15
-0,1
-0,05
0
0,05
0,1
0,15
0,2
12345
Driving time period
The index [ (theta 4-8 Hz + alpha 8-12 Hz) / beta 12-30 Hz] on-
target
Motorway Conventional road
Scheme 2. The parameter (θ+α/β) for EEG on-target periods (central derivation).
and so the value for period 1 (or baseline) is always 0 and
the values for the remaining driving periods (periods 2–5)
show the average of the variations produced with regard to
the baseline.
4.1. EEG central derivation (position C3-A2)
4.1.1. Spectral density on alpha frequency band
Scheme 1 shows mean values for the measurement of
spectral density on the alpha band while driving on the mo-
torway and on the conventional road, grouped by driving
time periods.
-0,3
-0,2
-0,1
0
0,1
0,2
0,3
0,4
12345
Driving time period
Alpha (8-12 Hz) on-target
Motorway Conventional Road
Scheme 3. Spectral density on alpha band for EEG on-target periods (occipital–parietal derivation).
The analysis of variance performed with this measurement
showed a not significant overall effect—but approaching
the critical value (near 0.05%)—for the interaction between
both independent variables (driving time period ×road
type) (F(3,36)=2.76, P=0.06). Nevertheless,
the effect of the road type variable within each driv-
ing period set was not statistically significant in any
case.
4.1.2. The parameter (θ+α/β)
Scheme 2 shows the mean values obtained for the mea-
surement of the relative parameter (θ+α/β) while driving

G.P. Cerezuela et al. / Accident Analysis and Prevention 36 (2004) 1045–1054 1051
-0,3
-0,25
-0,2
-0,15
-0,1
-0,05
0
0,05
0,1
0,15
12345
Driving time period
The index [ (theta 4-8 Hz + alpha 8-12 Hz) / beta 12-30 Hz]
on-target
Motorway Conventional Road
Scheme 4. The parameter (θ+α/β) for EEG on-target periods (occipital–parietal derivation).
on the motorway and on the conventional road, grouped by
driving time periods.
The analysis of variance performed with this measurement
showed a not significant overall effect—but approaching the
critical value—for the interaction between both indepen-
dent variables (driving time period×road type) (F(3,36)=
2.53, P=0.09). The effect of the road type variable within
each time period set was statistically significant for period
5(F(1,12)=6.57, P=0.02), this value being higher for
motorway driving than conventional road driving. Measured
by this indicator, this result points to a lower alertness level
on motorways than on conventional roads in driving period
5 (min 35–45).
4.2. EEG occipital–parietal derivation (position O2-P4)
4.2.1. Spectral density on alpha frequency band
Scheme 3 shows the mean values obtained for the mea-
surement of the spectral density on the alpha band while driv-
ing on the motorway and on the conventional road, grouped
by driving time periods.
The analysis of variance performed with this measurement
showed a not significant overall effect—but approaching the
critical value—for the road type variable (F(1,12)=3.83,
P=0.07) and the driving time period variable (F(3,36)=
2.28, P=0.09). For the former, the value of the measure-
ment was globally superior (but not statistically significant)
on the conventional road than on the motorway. Neverthe-
less, the interaction effect between the independent variables
(driving time period ×road type) was not statistically sig-
nificant.
4.2.2. The parameter (θ+α/β)
Scheme 4 shows the mean values obtained for the mea-
surement of the relative parameter (θ+α/β) while driving
on the motorway and on the conventional road, grouped by
driving time periods.
The analysis of variance performed with this measurement
showed a significant overall effect for the interaction be-
tween both independent variables (driving time period×road
type) (F(3,36)=2.83, P=0.05). The effects of the road
type variable in each condition in the driving time period
variable were significant for period 2 (F(1,12)=5.74, P=
0.03), this value being higher for conventional road driv-
ing. Measured by this indicator, this result points to a higher
alertness level on motorway than on conventional road at
driving period 2 (min 5–15).
5. Discussion
The primary objective of this paper was to study the inat-
tention state known as ‘highway hypnosis’. We have pre-
sented its signs, Wertheim’s hypothesis and the empirical
evidence obtained by the author in his laboratory experi-
ments. Since his experiments were conducted under condi-
tions described by the author as not applicable to an actual
driving environment, it is necessary to design a study under
real driving conditions to offer some additional empirical
evidence to Wertheim’s hypothesis on ‘highway hypnosis’.
This is the target to meet in our empirical research.
According to Wertheim, driving in environments with
moving stimuli (in relation to the driver) and following a

1052 G.P. Cerezuela et al. / Accident Analysis and Prevention 36 (2004) 1045–1054
highly predictable path—as is the case with motorways and
familiar roads—causes the predominant oculomotor control
type to switch, in such a way that visual information is finally
processed on the basis of mainly extra-retinal (internal) feed-
back, consequently taking into account very few external vi-
sual stimuli (intentive oculomotor control), automatization
being very high in the activity of the oculomotor system.
However, a driving environment where the motion of the
stimuli (in relation to the driver) follows a less predictable
trajectory—e.g. conventional roads—generates a predomi-
nantly attentive oculomotor control in the visual information
processing, this causing the driving to take place on the ba-
sis of retinal and extra-retinal information. This prompted
Wertheim to hypothesise that alertness—evidenced through
EEG activity—is lower on motorway and very familiar road
driving than while driving on conventional roads—e.g. main
roads and secondary roads—for those time periods during
which drivers visually track the trajectory of the moving
stimuli during their own motion (on-target ocular perfor-
mance periods, or—as we call it—‘look forward’).
Our empirical study attempted to find significant alertness
differences during the road eye-tracking periods (on-target
or ‘looking-forward’ periods) between motorway and con-
ventional road driving. Alertness was measured by means
of data relative to the activity of EEG slow bands which di-
rectly reflect the somnolence or drowsiness level. The driv-
ing time variable was divided into several time periods to
compare both road types not only globally but particularly
between similar time periods. If differences were found be-
tween both road types—as per the hypothesis—then they
would be found in the last driving periods.
The results partially confirm Wertheim’s hypothesis on
‘highway hypnosis’. All the on-target EEG measurements
analysed display a similar pattern with regard to the mean
values obtained for each of them. On the one hand, val-
ues were obtained indicating greater drowsiness on the con-
ventional road than on the motorway in the first driving
periods, which turned out to be statistically significant in
the measurement of the relative parameter (θ+α/β) in the
EEG occipital–parietal derivation for period 2 (min 5–15).
Although statistical significance was only obtained for this
measurement, this result comes to show that the starting
alertness level—measured by this indicator—was higher on
the motorway than on the conventional road for periods of
road eye-tracking (looking forward). Apparently opposite to
Wertheim’s hypothesis, this result could be explained by the
fact that during the first driving periods stimuli trajectory
predictability effects had not yet become apparent, since they
tend to show after a long drive. On the contrary, there might
be other factors accounting for the higher starting alertness
on motorways: external traffic conditions, speed, or even
differences in getting used to driving on such road types,
adaptation possibly being faster on conventional roads than
on motorways.
On the other hand, some values indicated greater drowsi-
ness on the motorway than on the conventional road at the
last driving periods. Such values were statistically signifi-
cant for the measurement of the relative parameter (θ+α/β)
in the EEG central derivation for period 5 (min 35–45). Al-
though statistical significance was only reached by this mea-
surement, the result shows that final alertness—measured
by this indicator—was lower on the motorway than on the
conventional road for periods of road eye-tracking (‘look-
ing forward’), which would back Wertheim’s ‘highway
hypnosis’ hypothesis. Since the analysis of EEG measure-
ments was conducted using only the time fragments corre-
sponding to on-target ocular performance, we could argue
that the difference found between both road types might be
due to the effect of the stimulus movement predictability
that takes place after a long driving period. Besides, a re-
lated factor that could also account for alertness differences
between both road types has to do with things in the visual
field unrelated to road definition, such as ‘other traffic’.
Hence, one of the most remarkable differences between
motorway driving and conventional road driving is the fact
that motorways are like one-way streets whereas conven-
tional roads are dual carriageways. Therefore, conventional
road drivers are visually exposed to vehicles coming in the
opposite direction on the lane right next to them, which pos-
sibly distorts environmental monotony and predictability.
However, this type of visual stimulation—vehicles coming
in the opposite direction right next to us—is not present
on motorways, which could contribute to explaining the
different ocular behaviour and alertness in both road types.
Nonetheless, we failed to find significant differences be-
tween motorway driving conditions and conventional road
driving conditions in the measurement of alpha activity in the
EEG occipital–parietal derivation, which is in contrast with
Wertheim’s laboratory result (Wertheim, 1978). In his exper-
iments, the author verified that EEG occipital alpha band ac-
tivity was significantly higher in the predictability condition
than in the non-predictability one during the on-target ocular
performance. We did not obtain this result in our study.
The fact that the only significant difference found sup-
porting the starting hypothesis was in an EEG measurement
recorded through central derivation and not in measurements
recorded through occipital–parietal derivation causes us to
ponder about the existence of other variables—besides the
predictability of the stimulus movement trajectory- account-
able for greater motorway drowsiness at the last driving
period. Most likely, such variables might have not been con-
trolled by the study due to the fact that the research is based
on an actual driving experience. Hence, despite the attempts
to thoroughly control the influence of potential variables on
the road type—time, weather, light, traffic (speed, incidents,
etc.), possible individual differences, and even the visual
tracking of the road (ruling out the periods in which this
did not occur)—some of them (or others) might have been
uncontrolled since the experimental control allowed by a
laboratory is difficult to replicate in real conditions. This
makes us think that some other factors may actually account
for the result relative to lower motorway alertness at the last

G.P. Cerezuela et al. / Accident Analysis and Prevention 36 (2004) 1045–1054 1053
driving period, since this difference was detected through
an EEG central derivation measurement, a derivation partic-
ularly sensitive in determining somnolence and drowsiness.
However, it might not be so sensitive in detecting the pre-
dominance of automatic control in road eye-tracking, for
which function the occipital–parietal derivation seems more
sensitive (where no such differences were found). Even
though significant alertness differences might exist between
motorway and conventional road as per the starting hypoth-
esis, this suggests the differences may not (or not only)
be due to the reason pointed out by Wertheim: the higher
stimulus trajectory predictability on motorways would pro-
duce a change in the predominant oculomotor control that
would lead to the development of automatization in the
eye musculature activity and the dependence of oculomotor
control on extra-retinal feedback at the expense of retinal
feedback.
Other not totally controlled factors might also be in-
volved, for instance, speed. In this regard, one of the
criteria observed in the selection of on-target EEG frag-
ments was that the speed on both the motorway and the
conventional road had to be 100–120km/h. This is why
all those EEG time fragments corresponding to a speed
other than that were ruled out. Yet, the greater speed vari-
ability found on the conventional road if compared to that
on the motorway cannot be overlooked. Once behaviour
adapts to this road type, the described circumstance could
in fact add to the higher alertness in the last driving period
of the conventional road in comparison with the motor-
way. As evidenced by some study (Tejero and Chóliz,
2002), variable speed produces higher alertness levels than
constant speed, the latter speed being more common on
motorways.
Finally, we believe Wertheim’s proposal is to be con-
sidered in a cautious way, particularly his statement about
motorway and conventional road driving adequately rep-
resenting the predictability and non-predictability of the
trajectory of moving stimuli, as per his laboratory experi-
ments. Even though we departed from such an assumption
for carrying out the empirical research, motorway and con-
ventional road driving are assumed to be far more complex
conditions than the laboratory tracking of the trajectory of
a visual stimulus moving in circles on a screen. However,
that is precisely the reason that caused Wertheim to ad-
dress the need to research in actual driving conditions to
ascertain whether or not alertness differences exist in the
on-target ocular performance between both road types. It
is also the reason why we adopted such an approach in
trying to provide Wertheim’s ‘highway hypnosis’ hypoth-
esis with empirical evidence. In our opinion, the empirical
research presented in this paper is a way to get closer to the
objective. We believe research should be further continued
in both real and simulated driving conditions. In addition,
other alertness indicators—not only psychophysiological
but also performance ones—would need to be evaluated,
for example, the detection of sudden relevant traffic events.
Acknowledgements
The authors wish to thank the corporation AUMAR
and the University Institute for Traffic and Road Safety
(INTRAS-University of Valencia, Spain), for providing the
required material and financial resources. Thank you to
Pilar Tejero for her research ideas.
References
Åkerstedt, T., Gillberg, M., 1990. Subjective and objective sleepiness in
the active individual. Int. J. Neurosci. 52, 29–37.
Bender, M.B., 1969. The oculomotor system and the alpha rhythm. In:
Evans, C.R., Mulholland, T.B. (Eds.), Attention in Neurophysiology.
Butterworth, London, pp. 304–309.
Brookhuis, K.A., De Waard, D., 1993. The use of psychophysiology to
assess driver status. Ergonomics 36 (9), 1099–1110.
Brown, I.D., 1991. Highway hypnosis: implications for road traffic
researchers and practitioners. In: Gale, A.G., Brown, I.D., Haslegrave,
C.M., Moorhead, I., Taylor, S. (Eds.), Vision in Vehicles-III. Elsevier,
North-Holland, pp. 459–466.
Brown, I.D., 1994. Driver fatigue. Human Factors 36 (2), 298–314.
Brown, I.D., 1997. Prospects for technological countermeasures against
driver fatigue. Accid. Anal. Prevent. 29 (4), 525–531.
Carpenter, R.H.S., 1977. Movements of the Eyes. Pion, London.
Cronbach, L.J., Furby, L., 1970. How we should measure “change”—or
should we? Psychol. Bull. 74, 68–80.
De Waard, D., Brookhuis, K.A., 1991. Assessing driver status: a
demonstration experiment on the road. Accid. Anal. Prevent. 23 (4),
297–307.
Gottman, J.M., 1990. Time-series analysis applied to physiological
data. In: Cacioppo, J.T., Tassinary, L.G. (Eds.), Principles of
Psychophysiology. University Press, Cambridge, pp. 754–774.
Hancock, P.A., Verwey, W.B., 1997. Fatigue, workload and adaptive driver
systems. Accid. Anal. Prevent. 29 (4), 495–506.
Heywood, S., 1972. Voluntary control of smooth eye movements and their
velocity. Nature 238, 408–410.
Johnson, L.C., Lubin, A., 1972. On planning psychophysiological
experiments: design, measurement and analysis. In: Greenfield, N.S.,
Sternbach, R.A. (Eds.), Handbook of Psychophysiology. Holt, Rinehart
& Winston, New York, pp. 125–158.
Kecklund, G., Åkerstedt, T., 1993. Sleepiness in long distance truck
driving: an ambulatory EEG study of night driving. Ergonomics 36 (9),
1007–1017.
Keller, E.L., 1977. The role of the brain-stem reticular formation in
eye movement control. In: Brooks, B.A., Bajandas, F.J. (Eds.), Eye
Movements, ARVO Symposium 1976. Plenum Press, New York,
pp. 105–126.
Kerr, J.S., 1991. Driving without attention mode (DWAM): a formalisation
of inattentive states in driving. In: Gale, A.G., Brown, I.D., Haslegrave,
C.M., Moorhead, I., Taylor, S. (Eds.), Vision in Vehicles-III. Elsevier,
North-Holland, pp. 473–479.
Mulholland, T.B., Evans, C.R., 1965. An unexpected artefact in the human
electroencephalogram concerning the alpha rhythm and the orientation
of the eyes. Nature 207, 36–37.
Mulholland, T.B., Peper, E., 1971. Occipital alpha and accommodative
vergence, pursuit tracking and fast eye movements. Psychophysiology
8, 556–575.
Murphy, B.J., Kowler, E., Steinman, R.M., 1975. Slow oculomotor control
in the presence of moving backgrounds. Vision Res. 15, 1263–1268.
Pastor, G., Chisvert, M., Monteagudo, M.J., 1999. El fenómeno de
la “hipnosis de la autopista”: origen y evolución histórica de su
conceptualización. Revista de Historia de la Psicolog´
ıa 20 (3–4), 83–94.

1054 G.P. Cerezuela et al. / Accident Analysis and Prevention 36 (2004) 1045–1054
Plotkin, W.B., 1976. On the self-regulation of the occipital alpha rhythm:
control strategies, states of consciousness and the role of physiological
feedback. J. Exp. Psychol.: Gen. 105 (1), 66–99.
Ray, W.J., 1990. The electrocortical system. In: Cacioppo, J.T.,
Tassinary, L.G. (Eds.), Principles of Psychophysiology. University
Press, Cambridge, pp. 385–412.
Rechtschaffen, A., Kales, A., 1968. A manual of standardized terminology,
techniques and scoring system for sleep stages of human subjects.
BIS/BRI, UCLA, Los Angeles.
Russell, D.W., 1990. The analysis of psychophysiological data:
multivariate approaches. In: Cacioppo, J.T., Tassinary, L.G. (Eds.),
Principles of Psychophysiology. University Press, Cambridge,
pp. 775–801.
Shor, R.E., Thackray, R.I., 1970. A program of research in ‘highway
hypnosis’: a preliminary report. Accid. Anal. Prevent. 2, 103–
109.
Stark, L., 1968. Neurological Control Systems: Studies in Bioengineering.
Plenum Press, New York.
Steinbach, M.J., 1976. Pursuing the perceptual rather than the retinal
stimulus. Vision Res. 16, 1371–1376.
Tejero, P., Chóliz, M., 2002. Driving on the motorway: the effect
of alternating speed on driver’s activation level and mental effort.
Ergonomics 45 (9), 605–618.
Thiffault, P., Bergeron, J., 2003. Monotony of road environment and
driver fatigue: a simulator study. Accid. Anal. Prevent. 35 (3), 381–
391.
Ulrich, G., 1990. Oculomotor activity and the alpha rhythm. Perceptual
Motor Skills 70, 1099–1104.
Vasey, M.W., Thayer, J.F., 1987. The continuing problem of false positives
in repeated measures ANOVA in psychophysiology: a multivariate
solution. Psychophysiology 24, 479–486.
Wertheim, A.H., 1978. Explaining highway hypnosis: experimental
evidence for the role of eye movements. Accid. Anal. Prevent. 10,
111–129.
Wertheim, A.H., 1991. Highway hypnosis: a theoretical analysis. In: Gale,
A.G., Brown, I.D., Haslegrave, C.M., Moorhead, I., Taylor, S. (Eds.),
Vision in Vehicles-III. Elsevier, North-Holland, pp. 467–472.
Wertheim, A.H., 1974. Oculomotor control and occipital alpha activity:
a review and an hypothesis. Acta Psychologica 38, 235–256.
Wertheim, A.H., 1981a. Occipital alpha activity as a measure of retinal
involvement in oculomotor control. Psychophysiology 18 (4), 432–439.
Wertheim, A.H., 1981b. On the relativity of perceived motion. Acta
Psychologica 48, 97–110.
Wilder, J., 1950. The law of initial values. Psychosomatic Med. 12, 392.
Wilder, J., 1965. Pitfalls in the methodology of the law of initial values.
Am. J. Psychother. 19, 577–584.
Williams, G.W., 1963. Highway hypnosis: an hypothesis. Int. J. Clin.
Exp. Hyp. XI (3), 143–151.
Williams, G.W., Shor, R.E., 1970. A historical note on highway hypnosis.
Accid. Anal. Prevent. 2, 223–225.
Zeevi, Y.Y., Peli, E., Stark, L., 1979. Study of eccentric fixation with
secondary visual feedback. J. Opt. Soc. Am. 69, 669–675.