Chem. Senses 24: 137-144,
1999
© Oxford University Press
The Effect of Odour Priming on Long Latency Visual Evoked Potentials of Matching and Mismatching Objects
Department of Clinical Neurology, University of Sheffield, Royal Hallamshire Hospital, Sheffield 1 Warwick Olfaction Research Group, University of Warwick, Warwick CV4 7AL 2 Quest International, Ashford, Kent TN24 0LT, UK
Corrrespondence to be sent to: Steve Van Toller, Warwick Olfaction Research Group, University of Warwick, Warwick CV4 7AL, UK
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussion and conclusionsReferences |
|---|
This study reports a cross-modal, olfactory/visual event related potential (ERP) using odours as olfactory primes. The results show a difference in the ERP waveform for the N400 waveform when a visual image does not match the priming odour. An N400 peak was produced for both the matched and mismatched conditions but the peaks were significantly more negative for the mismatched condition. By the use of non-food odours this study extends an earlier finding by Grigor, who, using the same ERP paradigm, obtained similar results for food odours and photographs of food.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion and conclusionsReferences |
|---|
It is difficult to establish the precise moment of receptor contact of an odorous stimulus and to control precise duration of stimulation. Whilst direct effects of odours on EEG have been reported (Klemm, et al., 1992;
Van Toller et al.,
1993), it is difficult to interpret EEG signals in the noisy electrical environment of
the cerebral cortex (Van Toller et al., 1992). Such
difficulties
have led
to the development of highly
technical equipment in a small number of specialist laboratories in order to study directly
olfactory/trigeminal generated chemosensory evoked potential (CEP) (Kobal and
Hummel,
1992). The
necessary conditions and controls for such studies have been presented by Evans et al. (1993) and Van Toller (1994). The standard
evoked
potential (EP)
paradigm requires averaging numerous presentations before the signal becomes discernible
against the
noisy cortical electrical background. Technical difficulties relating to olfactory research have
constrained the application of such studies to detailed systematic parametric studies using a
limited
number of substances: namely, vanilla, hydrogen sulphide and carbon dioxide. In order to extend
olfactory investigations into more psychological studies it is necessary to use alternative
methods. This
experiment concerns the use of one such technique called event related potential (ERP). The ERP
is a
well established technique in auditory and visual experiments where these two sensory systems
have an
advantage in that they allow very precise control of stimulus onset and duration.
The so-called ERP `oddball' study has the advantage of allowing the
investigation of
odours using olfactory primes. The task normally involves a subject counting the number of odd
stimuli
occurring during a series of standard presentations. Odour priming has also been used by Torii et
al. (1988) using an EEG technique called contingent negative
variation
(CNV)
which measures direct current baseline changes rather than the oscillating alternating current
displayed in
normal EEG waveforms. Torii's subjects were primed with an odour and were then
presented
with an initial stimulus (S1) which served as a signal for an impending second signal (S2) which
required the subject to make a response. Torii reported a post-S1 increase in the CNV baseline
between the period 500–1000 ms when jasmine was used as the priming odour and a
decrease
in the CNV baseline when lavender was used as the priming odour. In a carefully controlled
CNV
study using odour priming Lorig and Roberts (1990) reported that the
positive
and
negative CNV effects could be shown to match subject expectations. This suggested that
cognitive
factors influenced the results. It is interesting to note that four of the seven subjects used in the
Torii
study were said to be perfumers, which might suggest an influence of cognitive factors. These
studies
suggest that it would be of value to attempt evaluations of cognitive factors when a subject is
presented
with odours.
Earlier studies have reported semantic priming, which refers to the facilitation of a stimulus
by a
related preceding semantic stimulus (Myer and Schvaneveldt, 1971).
Semantic
priming is an ERP
paradigm which has been investigated since Kutas and Hillyard (1980, 1984) first reported it. In the
Kutas and Hillyard experiment subjects were primed with a sentence like, `The pizza was
too
hot to _ _ _'. The missing word at the end of the sentence was either `eat'
(congruous) or `cry' (incongruous). The authors founds that the N400 component
was
characterized as a negative deflection occurring ~400 ms after presentation of a word with a
larger
deflection following an incongruous word. These authors were also able to show that the N400
was
sensitive to the degree with which the target stimulus was in agreement with the prior semantic
context.
The greater the degree of incongruity the larger the N400 peak.
Subsequently the N400 component has been identified in a variety of visual priming tasks
using
verbal stimuli (Polich, 1985; Bentin, 1987),
non-verbal stimuli using faces (Barrett and Rugg, 1989,
1990a) and pictures (Barrett and Rugg, 1990b).
The
non-verbal findings suggest that the N400 peak
may not be as specifically related to linguistic stimuli as was initially suggested. Additionally an
N400
peak has also been identified for the auditory modality but it appears to have different amplitude
and
latency characteristics (Holcomb and Neville, 1990). This latter finding
suggests that the priming
processes may be different across modalities. Evidence of cross-modal priming has been reported
in
two recent studies. Pratorelli (1994) found evidence of cross-modal
semantic
priming
on the N400 peak using pictures and spoken words. Domalski et al. (1991)
found evidence of cross-modal transfer for semantic repetition priming between
auditory–verbal and visual–verbal stimuli.
A recent investigation using the cross-modal paradigm (Lorig et al.,
1995) examined the
effects of odour on visual ERP responses generated by an `oddball' paradigm.
Three
odour names were repeatedly presented and were preceded on 25% of the trials by a
non-matching
odour. Lorig and his colleagues found that the P300 waveform elicited was not differentially
related to
subjects' olfactory sensitivity. However, an increase in negativity of frontal electrodes
was found
when the subjects were required to match olfactory stimuli to labels. The authors suggested that
the
significant differences found for N100, P200 and P300 indicated that the neural processes used to
evaluate odour labels differed from those used to evaluate labels from visual stimuli. Lorig and
his
colleagues did not report on the N400 waveform.
An examination of the relevant literature shows that few investigations to date have
systematically
examined the effects of olfactory stimuli on human brain activity using a cross-modal paradigm.
The
study reported in this paper is the first to examine the indirect effects of odour on visual ERPs
using an
odour priming paradigm. The N400 waveform component was chosen because it has been
reported
to be reliably elicited using a variety of stimuli across different sensory modalities. As indicated
above
(Domalski et al., 1991; Pratorelli, 1994),
it
has also been found in cross-modal investigations.
In an earlier study Grigor (1995) used common food odours
as
primes.
These were argued to be more readily identifiable and therefore likely to generate neural
semantic
networks. N400 peaks are thereby elicited when paired with an incongruous (non-primed) target.
The
food smells preceded targets, which were photographs of foods. In some cases (the primed
condition)
the odour and the photograph consisted of the same food item. In others the odour and food
photograph were different. The study was designed as a cross-modal design using the visual and
olfactory senses. The main finding supported the prediction that the N400 component was greater
for
the incongruous (non-primed) than for the congruous (primed) condition.
The present study was designed to extend the earlier experiment by investigating the effects on long latency EPs (N400) using non-food odours and photographs of objects which were matched or mismatched in terms of the priming odour.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion and conclusionsReferences |
|---|
Subjects
All participants were volunteers recruited via poster, computer mail-shots and adverts from
within
the university and external colleges (a few external volunteers came forward following an
interview
given on the local BBC radio station). A total of 21 volunteers took part in the study of which 11
were
males and 10 were females.The mean age of the subjects was 25 years, with an age range of
20–35 years. All were right handed as assessed by the shortened version of the Edinburgh
Handedness Inventory (Oldfield, 1971). Subjects gave written consent for
the
study which had
received approval by the Ethical Committee of the Psychology Department. Each subject was
paid
(£4.00) for participating in the study. They were screened for suitability using an in-house
questionnaire relating to sleep habits, beverage consumption and medication. Subjects were also
asked
to confirm that they were not suffering from colds, sinus problems, asthma or hay fever on the
day of
testing.
Apparatus
The study used an olfactometer with a multi-input nozzle (White, 1991, 1997). The nozzle involves
an arrangement whereby separate odour lines are brought together in a cone which has a central
opening that is placed below the nares of subjects. A NeuroScan Inc 32 Syn-amps EEG
Amplifier
(model number 50835/N) was used to record the electrical activity from the cortex. Three Kodak
slide
projector carousels back-projected images onto a translucent screen. A white noise generator
provided
masking sounds via headphones worn by the subjects. Using an in-house program, an Apple
Macintosh 11SI computer synchronized the equipment and the presentations used during the
experiment. A schematic flow diagram showing presentation sequences of the stimuli are shown
in
Figure 1.
|
Olfactory stimuli
The olfactory stimuli were four supratheshold synthetic odorants—rose, lemon, grass and leather—provided by Quest International. The odours were diluted and judged to be isointense in a pilot study using 10 subjects. The rose, lemon and grass solutions were each diluted using equal volumes of dipropylene glycol. The leather odour was diluted 1/10. A total of 75 ml of each odour was placed in the olfactometer bottles. The odours were screened in a pilot study by subjects who rated them for familiarity, isointensity and match to the visual stimuli. The olfactometer presented each suprathreshold odorant in a stream of charcoal-filtered air at room temperature using a flow rate of 1 l/min. Presentation of the odours was for a period of 4 s. The control air and odour streams were delivered just below the subject's nose using a multi-input nozzle. Each odour was delivered via separate lines to the conical output nozzle which was placed below the subject's nares. During periods when an odour were not being presented a stream of humidified air at room temperature was delivered to the subjects via the nozzle. As mentioned earlier, odour deliveries were controlled by a Macintosh computer. Switching time for the odour delivery was fast relative to the total odour delivery time.
Visual stimuli
Each odour was represented by a series of photographs. These consisted of 35 professionally taken high-quality coloured slides presented via three Kodak carousel projectors. The slide presentations were controlled by the computer and were carefully synchronized with the priming odours. Each of the four priming odours had a correspond- ing set of eight slides. For six of the slides the odour matched the image and for two of the slides the odour did not match the image. For example, an odour-matched image for the smell of grass was a slide showing a picture of grass. An odour-mismatched image was a slide showing an image of a road surface. The mismatched images were selected to be visually similar in terms of either shape or size. The mismatched visual stimuli are presented in Table 1.
|
In an attempt to reduce boredom and novelty effects, and to ensure that subjects looked at the projected pictures, each view was taken from a different angle. All the presented slides were slightly different.
The slides were back projected onto a 160 x 109 cm translucent screen. Each slide was 80 x 50 cm when projected onto the screen. Subjects were seated 1.5 m in front of the screen with their chins placed on a chin rest and the odour nozzle just below their noses. To help standardize overall illuminance the slides were projected through a neutral density filter. To ensure that each subject was com- fortable the chin rest was adjusted each time. As shown in Figure 1., the slides were presented for a duration of 800 ms. A ratio of 75/25 matched/mismatched was used.
ERP recordings
The ERPs were recorded using a commercially produced electrode cap (Electro-cap, Eaton,
OH)
and a NeuroScan Inc, 32 channel Syn-amps machine, model no 50835/N (Neuroscience,
Warwick,
UK). The electrodes were referenced to linked mastoids using a forehead electrode as a ground.
All
impedances when recording were below 15 k
and typically below 7 k. The total
recording epoch was 1200 ms, with a 200 ms pre-stimulus and a post-stimulus period serving as
pre-stimulus and post-stimulus baselines. The A/D rate was 1000 Hz, leading to a sampling rate
of
1200. The bandpass was 0–20 Hz. ERP trials were recorded individually tagged in order
to
allow for later off-line aver- aging into matched and mismatched trials.
Procedure
Following completion of various questionnaires, the electrode cap was fitted to the subject with Omni-prep used to prepare the scalp sites and a commercial (SLE) electrode gel inserted into the individual electrode cups. The multi-input nozzle from the olfactometer was then adjusted to a position just below and in front of the nares of the subject. Headphones were carefully placed over the subject's ears, to ensure no alterations in electrode positions or impedance levels. The white noise generator was switched on and adjusted to a level comfortable for the subject but above a minimal level masking apparatus sounds. The olfactometer was located in a nearby closed fume cupboard. The experimenter was able to communicate with the subject via the headphones. They were then asked to read the following experimental instructions and encouraged to ask any questions.
Instructions to subjects
This experiment relies upon your breathing
technique. You are
required to breathe in through your nose and out through your mouth. The brass nozzle in front of
your
nose may or may not deliver an odour. Please breathe through your nose evenly, there is no need
to
sniff. Try to remain as relaxed as possible throughout the experiment. An orange light will appear
in
order to alert you to the presentation of an odour. At this time place your chin in the chin rest and
begin
breathing in through your nose and out through your mouth and try not to move your head from
side to
side. Following the odour a second red light will appear to alert you to the presentation of a slide
on the
screen in front of you, showing a colour slide which will either be related to the odour you have
just
received or bear no relation to the odour. You are not required to make an oral verbal judgement
during the experiment but simply make conscious decisions as to the relationship between the
two
stimuli. It does not matter if you recognize the odour but are unable to name it, simply
concentrate on
deciding if the two stimuli are related. Please try not to blink, swallow or make any motor
movement
following the second light until about 1 second after the slide has been switched off. Please
ignore the
video camera which is recording your eye movements and try to keep your eyes focused on the
centre
of the screen. The headphones you are wearing will produce white noise in order to reduce the
sounds
from the various pieces of equipment used to run the experiment. Three practice trials will be
given,
followed by a short break, please ask any questions or make any comments you feel are relevant.
There will be 32 experimental trials. Do you have any questions or require further information
before
you begin? Thank you for participating.
Subjects were instructed in the breathing technique and given the three warm-up practice trials. The breathing method was to prevent retronasal stimulation of the olfactory receptors. Once the initial phase had been satisfactorily completed, illumination level in the room was lowered and the computer programme and ERP recordings started. During the 32 experimental trials an orange warning light was illuminated for 1 s in order to alert subjects to the impending odour and the need for the required breathing pattern. An odour was presented for 4 s. A red light was then illuminated to warn the subject about an impending slide presentation. The warning lights helped the subjects reduce muscular movements during presentation of the stimuli. Subjects were asked to look at a fixation point on the screen. In order to do this a dot was present, other than during slide presentations, at the centre of the screen. Following a slide presentation, subjects were required to make a subvocal decision about the relationship, if any, between the odour and the slide. However, no oral or motor response was required during the experimental trials.
The EEG recordings were started 200 ms before the projection of the slide and lasted until 200 ms post-picture offset (see Figure 1. for details). Slides were projected for a total period of 800 ms. An intertrial interval of 40 s was used between each of the 32 trials in order to reduce response habituation. After the experimental trials had been completed subjects were presented with the same 32 stimuli in an identical order and asked say if the odour matched the slide. This procedure was carried out in order to check that the subjects had performed the task correctly.
Treatment of results
Off-line editing of post-recording data included digital filtering at 0.03 and 30 Hz (24
dB/octave).
Trials containing in excess of 75 µV were rejected. Trials containing eye move- ments
were
corrected using the algorithm of Semlisch et al. (1986). Analysis
was
initially
performed on a subject by subject basis on the NeuroScan Syn-amps by selecting the matched
and
mismatched ERP trials for each subject at each electrode site. These data were printed as two
grand
average waveforms (match and mismatch for each electrode for each subject; see Figure 2). The
original NeuroScan ASCII data was then imported into SPSS and the statistical analyses were
performed on an Apple Macintosh computer using the SPSS statistical package version 6.1.
|
A latency window was selected based on the hypothesis that the N400 component would be differentially affected by the matched and mismatched conditions. The window selected for the ERP component analysis was between 300 and 600 ms post-stimulus onset, or 500–800 ms from the start of the recording (see Figure 1.). The peak negativity values were selected from the 300 data points within this latency window.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion and conclusionsReferences |
|---|
The converted results were subjected to an SPSS repeated measures ANOVA with the factors being sex x electrode x condition. All significant results quoted are significant to at least P< 0.05. All reported ANOVA results are quoted using Greenhouse– Geisser tests for significance.
The waveforms were visually inspected and, in order to reduce the data to manageable proportions, 18 electrode sites were selected and used for the quantitative analysis. The selected electrodes were Fz, FT7, FT8, FC3, FC4, Cz, C3, C4, CP3, CP4, CPz, T3, T4, TP7, TP8, Pz, P3 and P4 (see Figure 2). The qualitative frequent and rare waveforms for each of the electrode are shown in Figure 3. Electrode grand means and standard deviations for the frequent and rare conditions are shown in Figure 4.
|
|
Overall there was a significant condition effect (F = 15.63; df 20,1; P = 0.001) for the N400 waveform, showing a significant difference for the 21 subjects used in this study between the frequent (matched) and the rare (mismatched) conditions. Electrodes reached significant levels (F = 2.44; df 17.30; P = 0.048). The condition x electrodes interaction showed a trend without quite reaching significance. Similarly, there was a near significant midline electrodes effect (F = 2.64; df 1,20; P = 0.081) and a near significant interaction of condition by electrodes at the mid line sites (F = 2.91; df 1,20; P = 0.06). Statistical testing of differences between the hemispheres showed a significant condition effect (F = 30.544; df 1,20; P = 0.002). There was a significant electrode effect (F = 3.58; df 1,20; P = 0.02), but no significant effect of hemisphere. An interaction between hemispheres and electrodes just reached significance (F = 2.65; df 1,20; P = 0.06). Overall, female subjects showed a significant condition effect (F = 8.45; df 1,10; P = 0.016). This was also the case for male subjects but it was more marginal (F = 5.59; df 9,1; P = 0.042). The males and female subjects showed no significant hemispheric or electrode effects and no significant interactions.
|
Discussion and conclusions |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion and conclusionsReferences |
|---|
The results showed that there was a difference in the ERP waveform for the N400 waveform when a visual image did not match the odour used as a prime. A negative peak with a latency of ~400 ms post-stimulus onset was produced for both matched and mismatched stimuli, with the latter condition being statistically significant. This study extends an earlier finding (Grigor, 1995
) using food odours. This
study, in common with earlier visual and auditory ERP experiments, found a significant
cross-modal
olfactory/ visual effect using an olfactory prime. In contrast to the previous reported work, no
overall
hemispheric effects were obtained and amplitude differences were greatest over the midline
electrodes
FZ, CZ, CPZ and PZ.
Both Lorig et al. (1993) and Grigor (1995) reported
increased negativity over frontal electrodes. Lorig and his collaborators found increased
negativity over
the left frontal electrodes, whereas Grigor reported increased negativity over the right frontal
electrodes.
Kutas and Hillyard (1980) reported in their seminal study that N400 was
maximal in
sites over the right hemisphere. However, the present experiment failed to find such an effect. In
the
study conducted by Lorig and his colleagues (1993) the odour was
provided
with a
verbal label. It could be argued that this would require left hemispheric processing, which may
account
for his findings. The earlier study by Grigor (1995) provided pictures of
matching or
mis-matching stimuli in re- sponse to the odours and is therefore more likely to require right
hemispheric
processing. An alternative explanation may be that both studies used food or food related odours.
Lorig
et al. (1993) noted that an incorrect label to an odour stimuli
should
result in a
large P300 component because it is dependent upon the subjective likelihood of an event and
therefore
should produced an oddball effect. Whilst we would agree with this proposal, it is also true that
both
Lorig's study and the present one require semantic decisions regarding the degree of
congruence between the odour and its subsequent label. This interpretation is supported by
Lorig's finding of increased negativity when subjects were required to match olfactory
stimuli to
visual labels. Lorig et al. (1993) reported a slow wave elicited
by an
odour. In
this study we did not measure the baseline changes and it is possible that slow wave changes
account
for the negativity reported here.
As has been noted earlier, the N400 is produced whenever a semantic decision is required but
it is
at its greatest ampli- tude when semantic incongruency is present. Decon et al. (1991) have suggested that what is critical for the N400 is not the
relationship between the
stimuli but
whether or not the stimuli are processed semantically. An odour presented without cues is a
stimulus
requiring a high level of semantic processing. Subjects when tested after the experiment found no
difficulty in identifying the odours and relating them to the photographs. Unlike Lorig, we did
not
attempt to assess the overall competence of our subjects but the odours were set at suprathreshold
levels and presumably readily able to produce a contextual set.
The reported procedure shows a valid method of evalu- ating the linguistic and associative
semantic
networks and their relationships to odours. Although it was not attempted in this study, in any
follow-up
study it would be of interest to assess the odour lexicons of the subjects (van Toller et
al.,
1992). It would also be interesting to extend the study by examining effects of
emotional
judgements on
the N400 waveforms of an ERP waveform.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion and conclusionsReferences |
|---|
Barrett, S. and Rugg, M. (1989) Event related potentials and the semantic matching of faces. Neuropsychologia, 27, 913–922.[Web of Science][Medline]
Barrett, S. and Rugg, M. (1990a) Event related potentials and the semantic matching of pictures. Brain Cogn., 14, 201–212.[Web of Science][Medline]
Barrett, S. and Rugg, M. (1990b) Event related potentials and the phonological matching of picture names. Brain Lang., 38, 424–437.[Web of Science][Medline]
Decon, D., Breton, F., Ritter, W. and Vaughan, H.G. (1991) The relationship between N2 and N400. Scalp distribution, stimulus probability and task relevance. Psychophysiology, 28, 185–200.[Web of Science][Medline]
Domalski, P. Smith, M.E. and Halgren, E. (1991) Cross modal repetition effects on the N4. Psychol. Sci., 2, 173–178.
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 and Neville, H. (1990) Auditory and visual semantic priming in lexical decision; a comparison using ERPs. Lang. Commun. Processes, 5, 281–312.
Klemm, W.R., Lutes, S.D., Hendrix, D.V. and Warrenburg, S. (1992) Topographical EEG maps of human response to odors. Chem. Senses,17 , 347–361.
Kobal, G. and Hummel, T. (1992) Olfactory evoked potential activity and hedonics. In Van Toller, S. and Dodd, G.H. (eds), Fragrance: the Psychology and Biology of Perfume. Elsevier Applied Science, London, pp. 175–194.
Kutas, M. and Hillyard, S. (1980) Reading
senseless sentences: brain potentials reflect semantic incongruity. Science, 207, 203–207.
Kutas, M. and Hillyard, S. (1984) Brain potentials during reading reflect word expectancy and semantic association. Nature, 307,161 –163.[Medline]
Lorig, T. (1992) Cognitive and non-cognitive effects of odour exposure: electrophysiological and behavioural evidence. In Van Toller, S. and Dodd, G.H. (eds), Fragrance: the Psychology and Biology of Perfume, 1st edn. Elsevier Applied Science, London, pp. 161–173.
Lorig, T. and Roberts, M. (1990) The
effects
of odors and expectancy on the contingent negative variation. Chem. Senses,15
, 537–
545.
Lorig, T., Mayer, T.S., Moore, F.H. and Warrenburg, S. (1993) Visual event-related potentials during odor labelling. Chem. Senses, 18,379 –387.
Lorig, T., Turner, J.M., Matia, C. and Warrenburg, S. (1995) The contingent negative variation in an odor labeling paradigm. Psychophysiology, 32, 393–398.[Web of Science][Medline]
Myer, D.E. and Schvaneveldt, R.W. (1971) Facilitation in recognising pairs of words. Evidence of a dependency between retrieval operations.J. Exp. Psychol. , 90, 227–234.[Web of Science][Medline]
Oldfield, R.C. (1971) The assessment and analysis of handedness. The Edinburgh Inventory. Neuropsychologia, 9, 97–113.[Web of Science][Medline]
Polich, J. (1985) Semantic categorisation and ERPs. Brain Lang., 26, 304–321.[Web of Science][Medline]
Pratorelli, 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]
Semlisch, Anderer, Shuster and Presslich (1986) A solution for reliable and valid reduction of ocular artifacts, applied to the P300 ERP.Psychophysiology , 23, 696–703.
Torii, S., Fukuda, H., Kanemoto, H., Miyanchi, R., Hamauzu, Y. andKawasaki, M. (1988) Contingent negative variation (CNV) and the psychological effects of odour. In Van Toller, S. and Dodd, G.H. (eds), Fragrance: the Psychology and Biology of Perfume. Chapman & Hall, London, pp. 107–120.
Van Toller, S. (1994) A caveat for the reporting of
chemosensory eveent related potentials. Chem. Senses, 19, 195.
Van Toller, S. Behan, J. Howells, P. Kendal-Reed, M. andRichardson, A. (1993) An analysis of human cortical EEG activity to
odours. Chem. Senses, 18, 1–16.
Van Toller, S., Hotson, S. and Kendal-Reed, M. (1992) The brain and the sense of smell: can we begin to make sense of cortical information after an odour has been received. In Van Toller, S. and Dodd, G.H. (eds), Fragrance: the Psychology and Biology of Perfume. Elsevier Applied Science, London.
White, T. (1991) A Comparison of Item and Order Processing in Olfactory and Verbal Short-term Memory. MSc Thesis, University of Oxford.
White, T. (1997) Short-term Memory for Olfactory Stimuli: Separate Store or Result of Recording? PhD Thesis, University of Warwick.
Accepted November 6, 1998
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||