Chem. Senses 24: 145-154,
1999
© Oxford University Press
Visual Event Related Potentials Modulated by Contextually Relevant and Irrelevant Olfactory Primes
Clinical Neurological Science, Room LF73B, General Hospital, Tremona Road, Southampton SO16 6YD 1 Quest International, Ashford, Kent TN24 0LT, UK
Correspondence to be sent to: Anne Richardson, Quest International, Ashford, Kent TN24 0LT, UK
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Visual evoked potentials were recorded from 16 scalp locations on 10 young subject during presentation of a series of high-quality photographs on a computer screen. The photographs consisted of equal numbers of pictures of fruit (citrus and con-citrus fruits), flowers (roses and other flowers) and objects (e.g. buildings, vehicles, animals etc.) Every picture was different in order to avoid repetition effects. The pictures were presented under four odour conditions: no odour, rose odour, jasmine odour and citrus odour. In order to keep the subject alert they were asked to make categorizing decisions for the visual stimuli (e.g. flower or fruit). No decision was required concerning the relationship between the visual stimulus and the odour. As expected, the N400 peak was more negative when the picture stimulus did not match the odour. It is hypothesized that the N400 peak can be used as a measure of relatedness of a sensory stimulus to a previous or on-going prime, irrespective of the mode of the stimuli.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
It is possible to record brain electrical activity associated with olfactory stimulation. The principal methods include either recording spontaneous EEG activity or recording event-related potentials (ERP) following presentation of an odour. Since Moncrieff (1962
) first reported
EEG
effects
due to odour stimulation, considerable advances have been made in both investigative techniques
and
theoretical exposition. Several researchers have reported direct effects of odours on EEG and
have
attempted to understand the underlying emotional and cognitive events of fragrance perception
(e.g.
Klemm et al., 1992; Van Toller et al., 1993). Lorig and Schwartz (1988a, 1989)
found that a particular odour (spiced apple) increased the number of waves in the theta
(4–7
Hz) band when compared with a resting baseline or with other odours (lavender and eucalyptus).
They
also found that a number of other odorants, most of which were food-related, reduced higher
frequency
EEG activity. The same group (e.g. Lorig and Schwartz, 1989) also
indicated,
however, that imagery of
food affected the EEG in a similar way to the food-related stimulus. These findings were
interpreted
as suggesting that the mechanism by which some odours decrease EEG arousal is cognitively
mediated
rather than direct. Klemm et al. (1992) measured EEG response
to a
selection
of pleasant and unpleasant odours, and found consistent responses to odours, across subjects, in
the
theta waveband. These were not clearly associated with perceived strength, degree of arousal or
pleasantness. Van Toller et al. (1993) investigated a diverse
range of
odours
and found an increasing alpha amplitude of cortical activity related to psychometric response,
irrespective of the nature of that response (i.e. intensity, pleasantness or familiarity). A more
recent
study by Brauchli et al. (1995) reported increases in alpha2
power
after
stimulation with an unpleasant odour, while no effect was found with a pleasant odour. These
studies
highlight the complexity of brain response to olfactory stimulation and the scope for interaction
with
other modalities.
Other measures of brain activity used in investigations of olfactory response include
chemosensory
(olfactory) event-related potentials (CSERPs) (e.g. Kobal and Hummel, 1988; 1992) or contingent
negative variation (CNV), which is DC, as opposed to the AC current of the EEG (e.g.
Torii et al., 1988). Both of these have shown olfactory responses modified
with
changing hedonics and
mood. Kobal and Hummel ( 1988;,1992) found
differences in olfactory and chemosensory evoked
potentials (OEPs and CEPs) to olfactory stimuli. They demonstrated that the latency and
magnitude of
the response peak was dependent on whether or not the odour presented was considered pleasant
or
unpleasant and whether it was presented to the left or to the right nostril. Ehrlichman (1986) had
previously also reported hemispheric differences in processing pleasant and unpleasant
odours. The experimental techniques employed nasal cannulae and a sophisticated olfactometer
which
warmed and humidified the 5.6 l/min airstream passing over the olfactory mucosa. In our
experience,
many subjects find such odour administration techniques uncomfortable.
The interpretation of observed effects of odour on measures of CNV have been debated by
Lorig
and Roberts (1990). They suggested that cognitive effects (such as
associations and
memories) can be involved in CNV changes. Torii et al. (1988)
studied the
effect of odour stimulation on CNV. They interpreted their results as indicating that an increase
in CNV
signalled subject stimulation and a decrease signalled relaxation. However, although Lorig and
Roberts (1990) found changes in the CNV similar to those reported by
Torii,
they also
found that comparable changes could be induced by expectations of odour qualities.
Another avenue of research has utilized the effect of odour on traditional visual and auditory
ERPs.
For example, Lorig (1991) noted changes in P300 amplitude, i.e. a
positive
peak
detected ~300 ms after presentation of the stimulus. He used a classic `oddball'
paradigm, in three different experiments, in which the subject was presented with a series of
stimuli (e.g.
different tones were presented in an auditory experiment or different words were presented
together
with `non-words' in a visual experiment). In their work 80% of the stimuli perhaps
were
common and 20% rare, and the subjects counted the rare occurrences. A difference in amplitude
of the
P300 was detected between the two conditions (ie. common or rare). In the third experiment
subjects
were presented with three pairs of odours and labels; ERPs were recorded to the visual stimulus.
In this
latter case subjects' odour thresholds were found to be significantly correlated with the
amplitude of the P300 peak but the data were not interrogated further.
Lorig et al. (1993), in their consideration of the use of
psychophysiological
measures for investigations of olfaction, expounded the use of traditional ERP paradigms evoked
with a
visual or auditory stimulus, modulated by odours. Again they used a classical
`oddball'
paradigm wherein an olfactory stimulus was administered and followed by a visual label. In the
most
frequently occurring trials (P = 0.75), the label correctly identified the odour stimulus.
In the
rarer trials, an incorrect label was presented (P = 0.25). ERPs were recorded to these
congruent and incongruent stimuli. A large positive component occurred ~300 ms following the
stimulus. The amplitude of this P300 wave has been found to be affected by the subjective
certainty that
the rare and frequent stimuli differ. This is consistent with other observations in association with
novel
stimuli or stimuli that contradict one's expectations and enhance P300 amplitudes (Pritchard 1981; Coles et al., 1990).
The concept of a priming effect has been investigated extensively, at both the behavioural
and
physiological levels, using linguistic (words) and non-linguistic (pictures and diagrams) stimuli.
Studies
have tested the hypothesis that a target stimulus is `primed' and the subsequent
recognition of the target is facilitated when it is preceded by a prime with which it has one or
more
features in common. This effect may be manifested, for example, by a faster reaction time to the
target.
It has also been found that an ERP component, N400, is sensitive to the congruence between the
target
stimulus and a previously established context (Pratarelli, 1994). For
example,
changes to the N400
have been reported by Kutas and Hillyard (1980) using sentences to
establish
context.
Subsequent investigations have focused on expanding the types of contextual information
that can
elicit priming effects. Of particular interest are experiments where cross-modal paradigms have
been
used. For instance, Barrett and Rugg (1989) followed previous semantic
experiments
with experiments of their own using semantic stimuli and then further experiments using
pictures. In the
second of their reported experiments using pairs of pictures of faces subjects were required to
press a
button to indicate whether the faces were congruent or incongruent. Their results showed that the
time
taken to press the button when faces were incongruent was slower and the ERPs more negative
than
their related counterparts. This negativity peaked at ~450 ms following presentation of the target.
Barrett and Rugg concluded that their N450 component was similar to the N4 (or N400) evoked
by
linguistic stimuli and that N4 effects could be obtained with a variety of meaningful stimuli so
long as an
associative relationship between primes and targets could be established.
Barrett et al. (1988) have offered a general explanation
which
suggests that
the class of N4 (or N400) ERPs reflect the processing of some aspect of a representational system
that
includes (but is not limited to) semantic information. It is possible that the N4 (or N400)
responses are
the result of a common conceptual processor that is indifferent to the mode of input or the
representational system accessed.
The development of an argument for a common conceptual memory system valid across
modes
of stimuli was continued by Pratarelli (Pratarelli, 1994), whose work
paired
colour photographs
with spoken words. Each of the 26 words and pictures appeared in both the matched and
non-matched
conditions. Therefore there was the potential for repetition priming in which subjects might be
able to
recall seeing a subset of the pictures or hearing a subset of the words. To compensate, the
pseudorandom order of trials was carefully checked to ensure that any one item was placed as far
as
possible from its second occurrence. The ERPs elicited to unrelated picture targets and unrelated
word
targets were more negative than their related counterparts and fit the characteristics of the N400
in all
but the initial phase of differential processing. Pratarelli concluded that the N400 is a member of
a class
of negativities that are evoked in response to stimuli that violate a previously established context.
Nigam et al. (1992) also compared the processing of line
drawings to the
processing of words. They found larger negativities (N400) to anomalous words and anomalous
pictures than to semantically congruous words and pictures. The authors reported no differences
in
amplitude, latency or scalp distribution between the negativities elicited by words and those
elicited by
pictures. They also concluded that these N400 effects reflect activity in an amodal, abstract
semantic
system.
One characteristic of many `oddball' experiments and also of relatedness
judgement
tasks (cf. Barrett and Rugg, 1990) is that subjects are required to make an
overt semantic decision.
Holcomb and McPherson (1994) tested whether the N400-like effects
were
present in
a picture-pair ERP experiment in which the task did not overtly require the subject to
semantically
compare the prime and the target. On each trial subjects were presented with two line drawings
(prime
and target). A total of 150 pairs were presented with one-third of the target pictures being related
to
the prior prime picture, one-third being unrelated to the prime picture and the remaining
one-third being
non-objects that were unrelated to the prime picture. The recorded ERPs varied with the type of
target
picture, showing a substantially larger negative peak, between 325 and 550 ms for unrelated than
related target pictures. The non-object pictures, which were designed to be the picture equivalent
of
pseudowords, also produced a large negative response that was significantly larger than the
negativity
elicited by unrelated pictures. Holcomb and McPherson (1994) concluded
that
their
results, and those of Barrett and Rugg (1990), supported the position
proposed by
Nigam et al. (1992) that the N400 reflects activity in an amodal,
abstract
semantic system.
A previous study investigating cross-modal effects with olfactory stimuli was reported by
Grigor
(1995). She used a series of common food odours as odour primes for a
series of
photographs of foods. Subjects were presented with an olfactory prime followed by a visual
target and
were then required to decide whether the target was congruent or incongruent with the previously
presented odour. The findings supported the prediction that the N400 component is more
negative in
the non-primed condition than in the primed condition.
The main objective of the present study was to extend the findings of Grigor (1995)
using odour stimuli as primes and measuring N400 amplitude in a visual evoked potential and,
using a
modified experimental procedure, to address procedural issues discussed earlier. Specifically, the
issue
raised by Holcomb and McPherson (1994) concerning the effect of
conscious
decisions
about the relatedness of the targets and primes was addressed. In order to eliminate this type of
judgement task and to minimize contamination of the priming effect with a conscious memory
task
subjects were not asked to make any decision concerning the relatedness of the target visual
stimulus
to the odour prime. However, in order to keep the subjects alert each subject was asked simply to
press one of two buttons on a hand-held box to indicate whether or not the picture belonged to a
designated category (e.g. flowers or fruits). The current study also avoids short duration odour
stimuli
by using an odour that is continuously present throughout a series of visual stimuli. In addition,
the
current design also eliminates the potential for repetition priming as discussed by Pratarelli (1994) by
using each visual stimulus only once for each subject. It was predicted that if the N400
effect is an amodal effect generated through non-conscious priming, and not dependent on
conscious
matching decisions, then the unrelated target pictures would elicit more negative N400s than
related
target pictures.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The objective of this work was to record visual ERPs during olfactory stimulation. The experiment was designed to manipulate the level of relevance of the visual stimuli to odour primes by using a series of pictures of objects with varying degrees of relevance to the odour primes.
Visual ERPs were measured in four separate odour conditions: no odour; rose odour; jasmine
odour; and citrus odour. Within each odour situation there were four different types of visual
stimuli: the
visual stimuli were congruent with the odour prime (e.g. rose odour with rose pictures) or
semi-congruent (e.g. rose odour with flower pictures); or they were incongruent to the odour
prime.
There were two levels of incongruent pictures: one class in which the picture could be related to
an
odour but not the same odour as the odour stimulus (e.g. rose odour with pictures of fruit); and a
second class with no such odour connotation (e.g. rose odour with pictures of a car or a building).
The
method employed for collection of the ERPs, where relevant, followed the principals of Evans et al. (1993).
Subjects
Participants were all volunteers recruited via advertisements on hospital and university notice boards. Two females and eight males aged between 21 and 23 took part in the experiment and were offered a small remuneration for their time. All were right-handed non-smokers and were screened for coryza and nasal allergy; all had bilateral nasal patency. None of the participants were taking part in any concurrent experimental trial. The recording procedure was explained verbally to the subjects who provided written consent prior to participation. Prior ethical approval had been obtained from the hospital ethical committee which operates in accordance with the Helsinki Declaration.
Apparatus
A simple olfactometer was employed to prepare and deliver the odour stimuli. Clean bottled air was further purified using charcoal filters and then switched by computer controlled solenoid valves between the different odour generators of the olfactometer as appropriate. Each odour generator consisted of a glass boat containing one of the odour stimuli. The air was passed over the odorant at a rate of 1 l/min and the resulting odour stream was presented directly to the subject at a sniffing port. Provision was also made to deliver clean air directly to the subjects at the appropriate times. The air was delivered to the subject via Teflon tubing to a facial mask.
During each phase of the experiment the appropriate odour-laden air was presented to the subject for 1 min and thereafter clean air, without odour, was delivered for 1 min. Computer control of the solenoid valves enabled precise timing of the odour changes. The computer also synchronized the timing of the presentation of the pictures, picture type (i.e. fruit or flower etc.) and marking of the EEG traces (to show the onset of each visual stimulus). Only data recorded during the perfume presentations were analysed. Previous experience had demonstrated that providing successive periods of odour and clean air reduced the effects of odour habituation in subjects. Debriefing of the subjects at the end of the recording period verified that the subjects were aware of the presence of odours intermittently throughout the trial period, and could still perceive the sample odorants with ease.
Olfactory stimuli
The olfactory stimuli were three pleasant synthetic odorants prepared by professional and
experienced perfumers at Quest International and were considered to be accurate representations
of
the true smell of roses, jasmin or citrus fruit. A simple olfactometer (described above) was
designed to
deliver odours significantly above sensory threshold during the period of recording visual evoked
potentials. The odours were judged to be iso-intense by a panel of experienced assessors,
following the
methods of Van Toller et al. (1993). The oils were placed in the
glass
olfactometer boats, as described previously, over which the air was passed during the
`odour
on' periods.
Visual stimuli
The visual stimuli were high quality (super VGA 256 colours) photographs from a sequence of 1170 projected onto a computer screen. The pictures were of three types: flowers, fruits or other unrelated objects such as buildings or animals. The flowers and fruits also had subcategories of roses and citrus fruits.
Each picture was unique; none of the subjects were presented with the same picture a second time. Examples of the pictures are shown in Figure 1.
|
Procedure
Subjects sat in a reclining chair in a small room (2 x 1.5 x 3.5 m), and were fitted with an electrode cap for recording EEG responses from the scalp and a facial mask through which the odours were delivered. The room was ventilated to avoid odour accumulation. The solenoid valves and other operating systems were located in the main laboratory outside the low-odour room and subjects were unable to hear the various operating sounds.
The subjects were asked to observe a series of pictures presented on a visual display unit. A picture was presented every 4 s and subjects were asked to classify them as fruit, flower or neither fruit nor flower. The subjects were instructed simply to observe the pictures on the screen and press the left hand button on a hand-held box when the picture was a fruit or a flower and the right hand button for any other object. They were not required to match the object to the odour nor to distinguish a stimulus class, e.g. citrus fruit. They were informed that odours would be presented from time to time but they were not required to identify the odour, and were asked to breathe normally and regularly throughout the testing period.
Pictures were presented in random order and individual pictures were not repeated. They were displayed for 2 s. The sequence of events during the EEG recording period is represented in Figure 2.
|
The experimental design splits the duration of measurements on a single subject into four distinct periods. The odour primes for each period are detailed in column 3 of Table 1. To avoid fatigue and adaptation to the odour prime the prime was alternated with filtered air every 1 min. Within each odour period an equal number of pictures from the five categories, namely rose, other flower, citrus, other fruit and neither fruit nor flower, were presented to each subject (see Table 1for details). A new picture was presented every 4 s and displayed for 2 s. Pictures were presented during both the presence and absence of an odour prime, but only those responses in the presence of an odour were analysed.
|
After the last experimental session, the electrodes were removed and the subjects participated in a short debriefing. Subjects reported no difficulty with the categorization task using the hand-held box and none was evident. The computer recorded their reaction times and response. These responses served as a check on attention and possible fatigue; in fact there were only 1% incorrect or absent responses.
The class (flower etc.) of each stimulus was also recorded by the computer with each picture so that EEG responses for each class could be averaged separately.
ERP recordings
Recordings were obtained using a fabric cap fitted with 16 electrodes in locations defined by
the
10–20 system, four central (Fz, Cz, Pz, Oz) and 12 lateral (F3, F4, C3, C4, P3, P4, T3,
T4,
T5, T6, O1, O2). Linked mastoid electrodes served as the reference. The electro-oculogram was
recorded from electrodes situated above the right eyebrow and on the outer canthus of the left
eye.
Electrode impedances were <5 k
, and signals were amplified using
Nihon–Kohden
EEG amplifiers with a bandwidth of 0.03–30 Hz. Signal acquisition was achieved using
NeuroScan analog-to-digital converters and software. Each trace was stored separately and
checked
off-line for correct subject response and other extraneous artefacts, e.g. eye blinks and noise
level.
Figure 2 illustrates the sequence of events during the recording periods. Recordings of the visual evoked potentials were obtained during the 10–20 min epochs in which the subject was exposed to no odour, or to rose, jasmin or citrus odours. Each EEG recording began 500 ms (–500) before the presentation of the visual stimulus (time 0). The N400 wave of interest occurred between 350 and 600 ms. Subjects generally recorded their categorizations on the hand-held box between 700 and 1000 ms (i.e. post-N400). Data acquisition ceased at 1540 ms, when the stimulus was replaced by a fixation cross on the screen for 2 s. During this period 512 data points per channel were collected at 4 ms intervals per point, thus allowing alias-free recording of signal frequencies up to 40 Hz.
Signal analysis
The EEG recording for each picture consisted of 16 scalp potentials, an eye movement monitor and the class of stimulus (flower, fruit or incongruent). After recording, each EEG trace was visually inspected by an experienced neurophysiologist for artefacts such as excessive noise, eye blinks or other eye movement. EEGs were rejected on these criteria. They were also rejected where responses made by the subjects concerning the picture class were either absent or incorrect.
The combinations of pictures and odours were classified as congruent, semi-congruent or incongruent as follows. For the rose odour, the rose pictures were considered congruent; other flowers were classed as semi-congruent and fruit as incongruent. For the jasmine perfume all flowers, except roses, were classed as congruent, roses as semi-congruent and fruit as incongruent. For the citrus odour, citrus pictures, other fruits and flowers were congruent, semi- congruent and incongruent respectively. Accepted signals were averaged according to the congruence of the picture with the prevailing odour. EEG traces were obtained for 40 pictures of each congruence category and averaged.
The N400 potential amplitude was defined as the negative peak occurring between 350 and 600 ms after the visual stimulus. The amplitude was taken as the mean of the peak amplitude with five recorded points on either side, ie. the mean of 11 x 4 ms epochs or the mean amplitude over a 44 ms period.
Statistical analysis
The data were analysed by ANOVA using the Greenhouse Geiser factor 
for those
terms
which included the electrode locations. The error structure was nested: subjects were split by
odour
period, which in turn was split by the presentation of the pictures. The responses were analysed
as
repeated measures on each subject (thus splitting the picture periods into smaller subplots). The
treatment structure for the odours and pictures was crossed.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Subjects were judged to have remained alert throughout the EEG sessions. Only 1% of the subject responses to the pictures (recorded by a button press) were incorrect or missing, and reaction times also remained consistently within a range of 720–770 ms.
Each subject was debriefed after recording was completed. They were asked to identify those odours that had been used during the recording session from a group of five stimuli. All of the subjects recognized the rose odour, 9/10 subjects recognized the jasmine odour and 7/10 subjects recognized the citrus odour with absolute certainty. In all other cases the odour was recognized as familiar but could not be identified as the same odour with certainty. Subjects were also asked to match the odours with a subset of the pictures. All quickly matched the rose odour with a rose picture, citrus with a picture of citrus fruit and jasmine odour with a picture of flowers, though not necessarily with pictures of jasmine. This confirmed that the experimenter's classification of congruent and incongruent stimuli matched that of the subjects.
The omnibus ANOVA treating all 16 electrodes as repeated measures and classifying pictures as congruent, semi-congruent and incongruent produced the following. Odour main effects (i.e. no interactions) were significant (P = 0.038) but the odour by electrode site interaction was insignificant. Both the congruence main effect and the congruence by electrode site interaction were significant (P = 0.042 and 0.014 respectively). On inspection (using principal component analysis; 75.04% of variance accounted for in the first two principal components) it was found that the central and frontal electrode groups and the occipital and parietal electrode groups formed two distinct clusters. As the central and frontal electrodes had the greatest signal-to-noise ratio a second analysis was performed on the six electrodes C3, C4, Cz, F3, F4 and Fz. Congruent pictures were significantly different from incongruent pictures for these electrodes, with a P-value of 0.040. It is notable that these electrodes are topographically adjacent to one another. All further analyses were conducted on an odour by odour basis since the main effects for odour and order of presentation could not be independently resolved.
Analysis of the N400 potential amplitude showed a significant picture by site interaction effect when citrus odour was present and an analogous result for the rose odour (P = 0.017 and 0.075 respectively; Table 2). As semi-congruent pictures are an intermediate step between congruent and incongruent pictures, the test statistics for picture type include comparisons of like with like diluting the effect of picture type. For this reason we have focused on the contrast between congruent and in congruent pictures (Table 3). In Table 3 the effect of picture is significant for jasmine odour (P = 0.002) and the P-values for the picture by site interactions for both citrus and rose odours have decreased (P = 0.009 and 0.016 respectively).
|
|
Tables 4–7 give the amplitudes for the N400 peak for each odour, picture and electrode averaged across all the subjects. Where there was no odour present and subjects viewed a series of randomly presented pictures there was no significant difference in the mean amplitudes calculated for the different picture types (Table 4).
|
|
When the subjects were primed with a rose odour and shown rose, flower and fruit pictures a significantly more negative amplitude was recorded for N400 evoked by the fruit pictures than by the rose pictures on all of the frontal electrodes (P <0.001). The flower pictures evoked an N400 peak with a less negative amplitude than that evoked by the fruit pictures but more negative than that evoked by rose pictures. This difference, for the rose and flower pictures, was significant on the two lateral electrodes (F3, P < 0.001; F4, P < 0.01). No significant differences were recorded from the central electrodes for different pictures in the presence of rose odour except C3 (rose versus flower pictures, P < 0.05; Table 5). Figure3 provides a representation of the resulting EEG trace for the Fz electrode.
|
|
When the priming odour was jasmine there was a significant difference between the N400 amplitudes evoked by the flower and fruit pictures (P < 0.01), and also a difference between the potentials evoked by the flower and citrus pictures (P < 0.05), across all six of the electrodes analysed. In every case the potential evoked by the congruent visual stimulus (i.e. flowers) was the least negative. Differences in the ERP response between the flower and rose pictures were also evoked from the F4, Fz, C3 and Cz electrodes (Table 6 and, for Fz, Figure 4).
|
|
When a citrus odour was presented to the subjects a similar pattern of differences for congruent and incongruent pictures was recorded for the F4 and Fz electrodes, but these differences were not statistically significant. Significant differences (P < 0.01) were detected from the C4 and Cz electrodes, reflecting a more positive peak with the congruent picture than was also observed with the central electrodes and the jasmine odour. This may reflect a later overlapping component not investigated in our current analysis and will be the subject of further study (Table 6 and, for Fz, Figure 5).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The results show a negative deflection in ERPs recorded between 350 and 600 ms after stimulation with a visual image. The amplitude of this negative peak was not affected by different types of pictures when no odour was present. However, when an odour was introduced, the amplitude of the negative peak was found to be related to the degree of congruity of the picture to the odour in nearly all cases.
The ERPs recorded in this study show a similar time latency to those identified as N400 in
priming
lexical decision tasks. Previous studies had investigated changes in the amplitude of this peak
related to
deviations in semantic expectancies (Kutas and Hillyard, 1980),
contradictions
in other linguistic and
non-linguistic expectancies (e.g. Barrett and Rugg, 1990; Nigam et al.,1992; Holcomb and
McPherson, 1994), and across visual and auditory modalities (Holcomb
and
Neville, 1990). The
present study, along with that of Grigor (1995), now extends this
proposition
to include
olfactory and visual modalities; furthermore, it provides evidence that this effect is not dependent
upon
matching decisions.
There are important differences between these experiments and those of Lorig et al.
(1993) and Grigor (1995), who showed ERPs
modified by olfactory
stimuli. Both Lorig and Grigor used `oddball' experimental designs, required
semantic
decisions to be made concerning the relatedness of the stimuli and used discrete odour stimuli
presented
for short intervals. Lorig et al. analysed the P300 potential and in addition described a
long
duration frontal negativity in association with odour primed stimuli. Similar frontal negativity
was not
detected in this study, so it could be speculated that the observed negativity was related to
semantic
processing.
The topographical distribution of the N400 observed in the current study showed a greater
negative amplitude in the anterior electrodes, with greater negativity on the left than on the right
hand
side. This distribution is similar to that observed by Holcomb and McPherson (1994)
but was not reported in the Barrett and Rugg studies (1990). These
differences
may
result from differences in the task that the subjects were required to carry out. In our study, as in
that of
Holcomb and McPherson (1994), subjects were given a task which did
not
involve a
memory decision as required in lexical decision tasks or in the visual/olfactory
`oddball'
designs of Lorig and Grigor described above. Any uncertainties arising from odour memory
effects
were also removed in the current study by providing a constant odour stimulus throughout each
recording period. It is possible that some aspects of the processing associated with the decision
task
are more strongly reflected by a different configuration of neural generators.
The object decision task engages a different set of preparatory and/or short-term memory
processes than those employed in the matching task used by Barrett and Rugg (1990),
and also by Lorig et al. (1993) and Grigor (1995). For
example, the maintenance of information concerning the prime through the prime/target interval
would
not seem as crucial in the object decision task where the appropriate response can be made
without
even dealing with the prime. Not withstanding these differences in topography, and possible
differences
in mental processing involved in these two different tasks, the N400 still demonstrates sensitivity
to the
congruity of paired sensory stimuli which appears to be valid across different modes of sensory
stimuli,
whether or not semantic processing is involved in the task.
These results support the hypothesis that visual–olfactory interaction may occur in the
frontal lobes but this is by no means supported unequivocably. Zatorre and Jones-Gotman (1991)
demonstrated reduction in human olfactory discrimination (but not odour detection) in
patients with temporal and frontal lobe lesions and identified the right orbito-frontal cortex as
particularly important. This was elegantly confirmed using positron emission tomography. One
might
speculate that primary visual processes occur in the posterior cortex, but subsequent processing
involves progressively more frontal regions and visual attention is dependent on the frontal lobe
acting
as an integrated unit. Precise localization of the source of late ERPs is probably an unrealistic
aim with
this protocol as the waveforms almost certainly represent activity in a distributed population of
neurons. However, further examination should indicate any lateral asymmetry.
To show the N400 phenomenon with olfactory priming stimuli for visual ERPs confirms N400 as a common event in higher sensory processing. Because the olfactory system is organized quite differently from the visual, auditory and somatosensory systems, whose input has to ascend the neuraxis from lower levels, intuitively one might not have expected the results that have been obtained in this and other studies. However, finding direct evidence that brain activity to a visual stimulus is altered in the presence of an odour, as in this current study, does indeed support the notion that N400 reflects neuronal activity during an amodal cognitive process.
|
Acknowledgments |
|---|
We are grateful to Dr H. Ian Philip for making the perfumed air dispenser, to Dr David McNulty for advice on the statistical ana- lysis, to Quest International for supplying perfumes and pictures, to Dr Tyler Lorig for his guidance and to the referees for their helpful comments.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Barrett, S.E. and Rugg, M.D. (1989) Event related potentials and the semantic matching of faces.Neuropsychologia, 27, 913–922.[Web of Science][Medline]
Barrett, S.E. and Rugg, M.D. (1990)Event-related potentials and the semantic matching of pictures. Brain Cogn., 14,201 –212.[Web of Science][Medline]
Barrett, S.E., Rugg, M.D. and Perrett, D.I. (1988) Event related potentials and the matching of familiar and unfamiliar faces. Neuropsychologia,26 , 105–117.[Web of Science][Medline]
Brauchli, P., Ruegg, P.B., Etzweiler, F. and Zeier, H. (1995) Electrocortical and autonomic alteration by administration of a pleasant and an
unpleasant odor. Chem. Senses, 20, 505–515.
Coles, M.G.H., Gratton, G. and Fabiani, M. (1990) Event related brain potentials. In Cacioppo, J. and Tassinary, L. (eds), Principles of Psycho- physiology, Physical, Social and Inferential Elements. Cambridge University Press, Cambridge, pp. 425– 455.
Ehrlichman, H. (1986) Hemispheric asymmetry and positive-negative effect. In Ottoson, D. (ed.), Duality and Unity of the Brain: Unified Functioning and Specialisation of the Hemispheres. Macmillan, London, pp. 194–206.
Evans, W.J., Kobal, G., Lorig, T.S. and Prah, J.D. (1993) Suggestions for collection and reporting of chemosensory (olfactory)event-related potentials. Chem. Senses,18, 751–756.
Grigor, J. (1995) Do the eyes see what the nose knows? An investigation of the effects of olfactory priming on visual event related potentials. Chem. Senses, 20, 163.
Holcomb, P.J. and McPherson, W.B. (1994) Event related brain potentials reflect semantic priming in an object decision task. Brain Cogn., 24,259 –276.[Web of Science][Medline]
Holcomb, P.J. and Neville, H.J. (1990) Auditory and visual priming in lexical decision: a comparison using event-related brain potentials.Lang. Cogn. Process. , 5, 281–312.
Klemm, W.R., Lutes, S.D., Hendrix, D.V. and Warrenburg, W. (1992) Topographical EEG maps of human responses to odors. Chem. Senses,17 , 347–361.
Kobal, G. and Hummel, C. (1988) Cerebral and chemosensory evoked potentials elicited by chemical stimulation of the human olfactory and respiratory mucosa. Electroenceph. Clin. Neurophysiol., 71, 241–250[Web of Science][Medline]
Kobal, G and Hummel, C. (1992) Olfactory evoked potential activity and hedonics. In Van Toller, S. and Dodd, G.H. (eds), Fragrance: The Psychology and Biology of Perfume. Elsevier Science, London, pp. 175–194.
Kutas, M., and Hillyard, S. (1980) Reading
senseless sentences: brain potentials reflect semantic incongruity. Science, 207, 203–207.
Lorig, T.S. (1991) Chemosensory modulation of visual and auditory event-related potentials. Proc. IEEE Engng Med. Biol., 13, 548–549.
Lorig, T.S. and Roberts, M. (1990) Odor
and
cognitive alteration of the contingent negative variation. Chem. Senses, 15, 537–545.
Lorig, T.S and Schwartz, G.E. (1988a) Brain and odor I. Alteration of human EEG by odor administration. Psychobiology, 16, 281–284.
Lorig, T.S and Schwartz, G.E. (1988b) EEG activity during the administration undetected odors. Chem. Senses 13, 710.
Lorig, T.S and Schwartz, G.E. (1989) EEG activity during relaxation and food imagery. Imagin., Cogn. Person., 8, 201–208.
Lorig, T.S., Mayer, T.S., Moore, F.H. and Warrenburg, S. (1993) Visual event related potentials during odour labelling. Chem. Senses, 18,379 –387.
Moncrieff, R.W. (1962) Effect of odours on EEG records. Perfum. Essen. Oil Rec., 53, 757– 760, 825–828.
Nigam, A., Hoffman, J.E. and Simons, R.F. (1992) N400 to semantically anomalous pictures and words. J. Cogn. Neurosci., 4, 15–22.
Pratarelli, M.E. (1994) Semantic processing of pictures and spoken words: evidence from event related brain potentials. Brain Cogn., 24, 137–157.[Web of Science][Medline]
Pritchard, W.S. (1981) Psychophysiology of the N300. Psychol. Bull., 89, 506–540.
Torii, S., Fukunda, H., Kanemoto, H., Miyanchi, R., Hamauzu, Y. andKawaski, M. (1988)Contingent negative variation (CNV) and the psychological effects of odours. In Van Toller, S. and Dodd, G.H. (eds), Perfumery: The Psychology and Biology of Fragrance. Chapman & Hall, London, pp. 107–120.
Van Toller, S., Behan, J.M., Howells, P., Kendall-Reed, M. andRichardson, A. (1993) An analysis of spontaneous human cortical EEG
activity
to odours. Chem. Senses, 18, 1–16.
Zatorre, R.J. and Jones-Gotman, M. (1991) Human olfactory discrimination after unilateral frontal or temporal lobactomy. Brain, 114,71 –84.
Zatorre, R.J., Jones-Gotman, M., Evans, A.C. and Meyer, E. (1991) Functional localisation and lateralization of human olfactory cortex.Nature , 360, 339–340.
Accepted November 6, 1998
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||