Available online at www.sciencedirect.

com

Epilepsy & Behavior 11 (2007) 466–470

www.elsevier.com/locate/yebeh

Case Report

Olfactory dysfunction in temporal lobe epilepsy: A case

of ictus-related parosmia
Sarah Jacek a, Richard J. Stevenson a,*
, Laurie A. Miller b,c

a
Department of Psychology, Macquarie University, NSW 2109, Australia
b
Neuropsychology Unit, Royal Prince Alfred Hospital, NSW 2050, Australia
c
Department of Medicine, University of Sydney, NSW 2006, Australia

Received 30 March 2007; revised 28 May 2007; accepted 29 May 2007

Available online 29 August 2007

Abstract

Olfactory abnormalities in temporal lobe epilepsy (TLE) usually involve either brief hallucinations prior to seizures or chronic impair-
ments in odor discrimination and identification. We describe the case of a man (B.C.) with TLE with an unusual presentation, an ictus-
related parosmia. B.C. reported distorted perception of odor quality and hedonics that could provoke nausea and gagging, typically at its
most extreme in the week or so following a seizure. Measures of B.C.’s olfactory functioning were obtained at stages of the ictal cycle
when parosmia symptoms were severe and when they had decreased. Unlike other parosmics, B.C.’s detection thresholds were always
normal, and unlike others with TLE, he evidenced little impairment in identification or discrimination. Testing during a period of more
severe parosmia suggested that B.C.’s experiences might be the result of hedonic changes. We argue this may be the effect of seizure activ-
ity on the amygdala, which is known to be involved in mediating emotive reactions to odors.
 2007 Elsevier Inc. All rights reserved.

Keywords: Temporal lobe epilepsy; Parosmia; Olfaction; Amygdala; Odor hedonics

1. Introduction may report blanket contamination of all odors (e.g., ‘‘all

Olfactory abnormalities may co-occur with temporal lobe
epilepsy (TLE). TLE may cause brief olfactory auras prior to
seizures [1,2]. Chronic disturbances may also occur as a
result of structural brain damage [3], tumors [4], or interictal
epileptiform discharge [5]. The effect of TLE on olfaction has
been explored phenomenologically and experimentally. The
former approach focused solely on olfactory auras [1]. The
latter approach revealed that patients with TLE can detect
odors normally, but have impairments in a variety of
higher-level olfactory functions [6–8].
An olfactory disturbance that has not been detailed
before in TLE is parosmia. Parosmia is usually a chronic
condition [9] in which odor quality and hedonics (i.e., the
affective reaction to odors) are distorted [10–13]. Patients
odors smell artificial’’ [9,11]) or describe each odor as hav-
ing a unique but distorted smell (e.g., ‘‘strawberries smell
like bleach’’ [9]). The onset of parosmia has been associated
with upper respiratory tract infection, head trauma, and
aging [14], and parosmics often have impaired detection
thresholds [10]. In this report, we describe an unusual
ictus-related parosmia in a patient with TLE. The pattern
of presentation revealed by the olfactory testing detailed
below is both different from olfactory abnormalities noted
before for TLE and different from that of patients with typ-
ical parosmia.
2. Method
2.1. Case report

B.C., a 29-year-old, right-handed man, was unable to work during the

*
Corresponding author. Fax: +61 2 98508062. testing period because of memory problems caused by seizure activity.
E-mail address: richard.stevenson@psy.mq.edu.au (R.J. Stevenson). B.C. had completed 11 years of education. He had an unremarkable olfac-

1525-5050/$ - see front matter  2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.yebeh.2007.05.016
 Case Report / Epilepsy & Behavior 11 (2007) 466–470
tory and medical history until the onset of epilepsy around the age of 26.
The etiology of his epilepsy was unknown. He experienced simple and
complex partial seizures, both diurnal and nocturnal, which often general-
ized. These seizures typically clustered into discrete bouts, which occurred
every 4 to 6 weeks. This diagnosis was based on both ambulatory and
video/EEG monitoring, which captured 28 seizures. The focus was the
right temporal lobe as indicated by EEG, fluorodeoxyglucose positron
emission tomography, and clinical observation (an aura of left arm numb-
ness). MRI was inconclusive and indicated subtle left and right hippocam-
pal atrophy. He reported smoking for 12 pack-years and was a smoker at
the time of testing. B.C. was tested by us on six occasions; on the first three
occasions, he was taking clobazam (20 mg), levetiracetam (2000 mg), and
sodium valproate (800 mg) daily, but was drug free on the three remaining
sessions.
For several days after a bout of seizures, B.C. reported that he could
detect odors but they ‘‘did not smell normal.’’ Typically, each odor had
its own unique, but altered smell. This was consistent each time the change
occurred. Food flavors, which involve a significant olfactory input via ret-
ronasal olfaction, were reported as ‘‘nothing like they are meant to taste.’’
B.C. stated that meat, garlic, and onion odors were especially foul, making
him gag on occasions. Food intake was reduced to one meal per day while
the parosmia was severe. The parosmia usually diminished across the per-
iod separating bouts of seizures.
2.2. Healthy control participants
Six male control participants matched to B.C. with respect to age, edu-
cation, and smoking history were recruited from the community. Exclu-
sion criteria included history of epilepsy or neurological problems, nasal
surgery, and illness. The scores for controls on the National Adult Read-
ing Test (NART) [15] did not significantly differ from B.C.’s scores.
2.3. Procedure
B.C.’s parosmia was explored in two ways: (1) His olfactory abilities
were assessed over the ictal cycle, to identify any olfactory impairment
associated with his parosmia. (2) His parosmia was also explored in depth
when it was reportedly severe.
2.3.1. Sessions 1–3 (B.C. only)
The following sets of tests were repeated on three occasions: 4 weeks
postictally (mild parosmia), 4 days postictally (severe parosmia), and 2
weeks postictally (mild parosmia). Birhinal olfactory threshold was mea-
sured using a forced-choice ascending staircase procedure with butanol
[16]. Odor discrimination was assessed using a series of triangle tests
(two identical distracters and one target). Four different types of stimuli
were presented at each session and repeated in subsequent sessions: five
tests using familiar odors, five using unfamiliar odors, five using unfamil-
iar odors where either the target or distracters differed in intensity (making
the task easier), and five using visual stimuli (ambiguous amoeboid
shapes) to detect any generalized deficit (see [17,18] for details). The Digits
Backwards subtest from the Wechsler Adult Intelligence Scale—Third
Edition (WAIS-III) [19] was used as a measure of working memory [20].
Familiar odor identification, using a recognition task, was assessed using
three different odor sets, one per session.
2.3.2. Sessions 4 and 5 (B.C. and controls)
Sessions 4 and 5 were completed on 2 consecutive days, with a 24-hour
interval (for B.C. these were completed 2 weeks postictally, severe paros-
mia reported). In the first of these sessions, B.C. was given the odor name
acquisition test to determine whether his ability to perceive odor quality
was stable, as instability could be responsible for his reported changes
in odor perception. In this task he was required to learn and recall the
names of five unfamiliar odors. He was given two learning blocks, each
of which comprised all five odors being presented with the corresponding
names. Following this, six feedback blocks were completed. Following a
delay of 30 minutes, he was again presented with all five odors and
467
required to provide the names. A near-identical procedure was then under-
taken as a visual analog, in which B.C. was required to learn and recall
shapes paired with names. This ensured that any observed odor deficit
was not a consequence of a generalized impairment in name learning or
perceptual stability. During the 30-minute delay for the visual task, he
was asked to rate the similarity of a group of unlabeled odors containing
meat, garlic, onion, and foul smells (fecal and rotting fish), hereafter
known as the meaty/foul group. This was included to determine whether
these odors were perceived (hedonically and qualitatively) in a similar
manner. In this task, a large, flat 9 · 9-cell grid was laid out in front of
him. He was asked to smell each odor and to place it on the grid so as
to reflect how similar it was to the other odors (i.e., close = similar, dis-
tant = dissimilar). He resmelled and repositioned odors as needed.
In the second of these two sessions, B.C. was given recall and recogni-
tion trials of the odor and shape name learning tasks. The final group of
tasks required similarity-based judgments for: (1) labeled odors, (2) unla-
beled odors, (3) labeled pictures, and (d) unlabeled pictures. This was
included to explore whether his parosmia might be related to impoverished
odor naming, as misnaming, which should differ between labeled and unla-
beled conditions, can affect odor hedonics and quality [21–23]. For the
labeled odor condition, B.C. was again given the 9 · 9-cell grid. He was
presented with eight familiar odors, one by one in a set order, and had
to place these onto the grid to reflect their similarity (see above). As each
was presented, a label was also provided. The procedure for the unlabeled
odor condition was identical except no labels were provided. The labeled
and unlabeled picture tasks were identical, but with pictures being placed
face down on the grid.
The battery order for controls was as follows: The first session included
threshold, odor naming, the learning and feedback trials for odors and
shapes, and the meaty/foul similarity test. The second session comprised
the recall and recognition phases for odors and shape naming, followed
by the labeled and unlabeled odors and pictures.
2.3.3. Session 6 (B.C. only)
In this session (4 weeks postictally, mild parosmia reported), B.C. com-
pleted the SIT [24], facilitating comparison of his odor recognition abilities
with those of a normative sample.
2.4. Analysis
Across all tasks, B.C. was compared with controls using modified t
tests [25], with alpha set by Bonferroni adjustment.
3. Results
3.1. Sessions 1–3
B.C.’s threshold performance was normal across ses-
sions 1–3 (scores = 5, 6, and 8 respectively) relative to both
controls (mean = 6.7, SD = 1.0) and Cain’s [16] normative
sample. On the discrimination tests, B.C.’s performance
was compared with that of a sample drawn from two pre-
vious studies [17,18]. Compared with controls, B.C.’s per-
formance in discriminating unfamiliar odors was normal,
whereas his ability to discriminate familiar odors was sig-
nificantly impaired in session 3 (t(8) = 5.86, P < 0.005)
(see Table 1). Note, however, that B.C.’s discrimination
of familiar odors was perfect during the time in which his
parosmia was reportedly severe (session 2). B.C.’s intensity,
quality, and visual discrimination was normal, as was his
working memory (Digits Backwards subtest). B.C.’s odor
recognition performance (scores = 3.5, 3, and 4.5, respec-
468 Case Report / Epilepsy & Behavior 11 (2007) 466–470

Table 1
Olfactory discrimination scores for controls and B.C.
Test B.C
a
Controls Session 1 Session 2 Session 3
N Mean (SD) score Score t Score t Score t
Quality unfamiliar 10 2.6 (1.1) 2 0.51 4 1.18 4 1.18
Quality familiar 9 4.9 (0.3) 4 2.78 5 0.31 3 5.86b
Quality/intensity 11 3.8 (1.3) 4 1.42 2 1.28 2 1.28
Visual analog 8 4.9 (0.4) 4 2.08 4 2.08 4 2.08
a
Male age-matched control data obtained from [17,18].
b
Significant at the 0.017 level.

tively) did not differ from that of controls (means
(SD) = 4.5 (0.5), 4.8 (0.3), and 3.7 (1.1), respectively).
3.2. Sessions 4–6
B.C.’s scores on the odor and shape name learning tasks
did not differ significantly from those of controls on any
phase of either task. For the similarity data, the smallest
rectangular area of cells on the grid accommodating all
the stimuli (as they were placed by the participant), was cal-
culated for each participant and condition. For example, if
all the odors were placed as close as possible to each other
on the grid, they could be enclosed within a 3 · 3-cell rect-
angle (i.e., within an area of 9 cells). However, if odors
were maximally dispersed across the grid, with odors
placed in the corners, they could only be enclosed by a
9 · 9-cell rectangle (i.e., within an area of 81 cells).
The raw scores (see Fig. 1) suggest that B.C. tended to
cluster all stimuli (odors and pictures), relative to controls.
This is arguably a consequence of his placement ‘‘style,’’ as
the controls also varied widely in this respect. However,
our primary interest was B.C.’s relative performance across
the four tasks, that is, how similarity judgments were
affected by labeling and modality. To address this, and to
reduce variability attributed to different placement styles,
the ratios of labeled to unlabeled odors were calculated
and divided by the ratios of labeled to unlabeled pictures.
Ratios for controls were then averaged and compared with
B.C.’s score. This revealed a significant difference in ratios
between B.C. and controls (t(5) = 25.14, P < 0.01). B.C.
regarded the unlabeled odors as far more dissimilar, rela-
tive to his performance on the other tasks. To confirm this,
we conducted additional tests on the ratios for labeled/
unlabeled odors (B.C. vs controls) and for labeled/unla-
beled pictures (BC vs controls). There was no significant
difference for pictures, but B.C. regarded the unlabeled
odors as more dissimilar than the labeled odors, relative
to controls (t(5) = 5.89, P < 0.01).
The same method was used for the second similarity
task. Unlabeled meaty/foul odors were compared with

81

Controls
72 BC

63
Area (max = 81)

54

45

36

27

18

9
Picture-label Picture-non label Odor-label Odor-no label Meaty/foul-no label

Fig. 1. Mean areas (larger areas representing greater dissimilarity) and SEM for controls and scores for B.C., for the five similarity tests.
 Case Report / Epilepsy & Behavior 11 (2007) 466–470
the other unlabeled odors. The unlabeled meaty/foul set,
relative to the unlabeled set, was indeed different for B.C.
relative to controls (t(5) = 3.52, P < 0.02). Fig. 1 shows
that B.C. clustered the unlabeled meaty/foul odors more
than the other unlabeled odor set, suggesting that the
meaty/foul odors were especially similar to him. Finally,
for the SIT, using Australian norms, his corrected score
was 34/40. This placed him in the normosmic range, at
the 18th percentile.
4. Discussion
Previous reports of abnormal olfaction in TLE have
either featured ictus-related auras or revealed that TLE
can chronically affect higher olfactory functions. The case
described here is unlike both of these previous presenta-
tions, as well as unlike previous reports of parosmia.
Although B.C. did not exhibit abnormalities in odor iden-
tification, there was some evidence of impoverished dis-
crimination of familiar odors. However, this did not co-
vary with B.C.’s parosmia; indeed, his discrimination of
familiar odors was perfect when tested during a reportedly
severely parosmic period. Moreover, his performance on
the SIT was normal, and he did not demonstrate any
abnormality in threshold, as do other parosmic patients,
even when tested during a severely parosmic period. Need-
less to say, our control participants were not repeatedly
tested like B.C. was for discrimination, identification, and
threshold. This might mask more subtle performance
changes. Nonetheless, our findings still suggest that there
were no gross changes in B.C.’s olfactory system when
his parosmia was reportedly severe.
When B.C.’s perception of odor quality and hedonics
was tested during a period of severe parosmia, his perfor-
mance differed in certain ways from that of controls. The
absence of labels led B.C. to judge an odor set as far more
dissimilar relative to a matched but different set of labeled
odors. This was not the case for his judgment of labeled
and unlabeled pictures, suggesting that the odor labels
allowed him to normalize judgments of similarity. B.C.’s
perception of odor quality was clearly stable, both within
a brief period (30 minutes) and over a longer period
(24 hours), as he could attach labels to odors as consis-
tently as controls (as with the visual control task). Simi-
larly, his discriminative performance over sessions 1–3
suggests qualitative stability. Finally, relative to controls,
B.C. judged the unlabeled foul and meaty odors as more
similar than the other unlabeled odor set, suggesting he
found the foul and meaty stimuli to be very similar.
Parosmia has, with good reason, been considered a
condition in which abnormalities of the receptors or olfac-
tory epithelium are the likely cause [14]. This type of
explanation appears to be a poor fit for B.C. First, his
parosmia appears to be related to seizure activity, suggest-
ing a central locus. Second, he demonstrates normal SIT
and threshold scores and is broadly intact on other mea-
sures of olfactory function. Threshold and other tasks
469
would likely be impaired if there was peripheral damage,
as with most cases of parosmia. Third, the effect of verbal
labels appears to normalize his similarity judgments. This
might not be expected if the defect was exclusively
peripheral.
If B.C.’s parosmia is centrally based, what might be its
cause? One possibility is that rather than resulting from a
distortion of odor quality, the parosmia may be due to
changes in odor hedonics alone. There are several reasons
for this assertion. First, B.C. clustered the unlabeled foul
and meaty odors much more tightly than the other set of
unlabeled odors, indicative of a judgment based on hedon-
ics. Indeed, olfactory multidimensional scaling research
reveals a significant hedonic dimension to similarity judg-
ments [26]. Second, the scatter of the unlabeled set relative
to the labeled set can be accounted for in a similar manner.
This is because measures that might be expected to be sen-
sitive to abnormal quality perception, such as odor identi-
fication, discrimination, and odor name learning were all
broadly intact, implying that labels normalize hedonics,
not quality. Third, the intensity of B.C.’s reported hedonic
response to certain odors (i.e., nausea and gagging) is con-
sistent with a profound change in his emotive reaction to
smells. Taken together with recent neuroimaging findings
that suggest that the amygdala is critical to hedonic pro-
cessing [27,28], especially for unpleasant smells, we tenta-
tively suggest that B.C.’s seizure activity may temporarily
disrupt olfactory affective processing in his amygdala,
altering odor hedonics.
Acknowledgments
The authors thank B.C., the control participants, and
Dr. A. Mohamed for their assistance in conducting this
study.
References
[1] Acharya V, Acharya J, Luders H. Olfactory epileptic auras. Neurol-
ogy 1998;51:56–61.
[2] Chen C, Shih YH, Yen DJ, Lirng JF, Guo YC, Yu HY. Olfactory
auras in patients with temporal lobe epilepsy. Epilepsia
2003;44:257–60.
[3] Carroll B, Richardson JT, Thompson P. Olfactory information
processing and temporal lobe epilepsy. Brain Cogn 1993;22:230–43.
[4] Aldenkamp AP, Krom MD, Reijs R. Newer antiepileptic drugs and
cognitive issues. Epilepsia 2003;44:21–9.
[5] Grant AC. Interictal perceptual function in epilepsy. Epilepsy Behav
2005;6:511–9.
[6] Eskenazi B, Cain WS, Novelly RA, Mattson R. Odor perception in
temporal lobe epilepsy patients with and without temporal lobec-
tomy. Neuropsychologia 1986;24:55–62.
[7] Kohler CG, Moberg PJ, Gur RE, O’Connor MJ, Sperling MR, Doty
RL. Olfactory dysfunction in schizophrenia and temporal lobe
epilepsy. Neuropsychiatry Neuropsychol Behav Neurol 2001;14:83–8.
[8] Savic I, Bookheimer SY, Fried I, Engel J. Olfactory bedside test: a
simple approach to identify temporo-orbitofrontal dysfunction. Arch
Neurol 1997;54:162–8.
[9] Leopold D, Meyerrose G. Diagnosis and treatment of distorted
olfactory perception. In: Kurihara K, Suzuki N, Ogawa H, editors.
Olfaction and taste XI. Sapporo: Springer-Verlag; 1993. p. 618–22.
470 Case Report / Epilepsy & Behavior 11 (2007) 466–470
[10] Bonfils P, Avan P, Faulcon P, Malinvaud D. Distorted odorant
perception: analysis of a series of 56 patients with parosmia. Arch
Otolaryngol Head Neck Surg 2005;131:107–12.
[11] Frasnelli J, Landis BN, Heilmann S, Hauswald B, Huttenbrink KB,
Lacroix JS. Clinical presentation of qualitative olfactory dysfunction.
Eur Arch Otorhinolaryngol 2004;261:411–5.
[12] Landis BN, Frasnelli J, Hummel T. Euosmia: A rare form of
parosmia. Acta Otolaryngol 2006;126:101–3.
[13] Kaufman MD, Lassiter KRL, Shenoy BV. Paroxysmal unilateral
dysosmia: a cured patient. Ann Neurol 1988;24:45–51.
[14] Leopold D. Distortion of olfactory perception: diagnosis and
treatment. Chem Senses 2002;27:611–5.
[15] Nelson HE, Willison J. National Adult Reading Test (NART): test
manual. 2nd ed. Windsor: NFER Nelson; 1991.
[16] Cain WS. Testing olfaction in a clinical setting. Ear Nose Throat J
1989;68:316–28.
[17] Stevenson RJ, Mahmut M, Sundqvist N. Age related changes in odor
discrimination. Dev Psychol 2007;43:253–60.
[18] Stevenson RJ, Sundqvist N, Mahmut M. Age related changes in
discrimination of unfamilar odors. Percept Psychophys
2007;69:185–92.
[19] Wechsler D. Wechsler Adult Intelligence Scale administration and
scoring manual. 3rd ed. San Antonio, TX: Psychological Corp; 1997.
[20] Kareken DA, Mosnik DM, Doty RL, Dzemidzic M, Hutchins GD.
Functional anatomy of human odor sensation, discrimination, and
identification in health and aging. Neuropsychology 2003;17:482–95.
[21] Distel H, Hudson R. Judgement of odor intensity is influenced by
subjects’ knowledge of the odor source. Chem Senses
2001;26:247–51.
[22] Morrot G, Brochet F, Dubourdieu D. The color of odors. Brain Lang
2001;79:309–20.
[23] Herz RS. The effect of verbal context on olfactory perception. J Exp
Psychol Gen 2003;132:595–606.
[24] Doty RL. The Smell Identification Test administration manual. 3rd
ed. Philadelphia: Sensonics Inclusive; 1995.
[25] Crawford JR, Howell DC. Comparing an individual’s test score
against norms derived from small samples. Clin Neuropsychol
1998;12:482–6.
[26] Schiffman SS, Robinson DE, Erickson RP. Multidimensional scaling
of odorants: examination of psychological and physiochemical
dimensions. Chem Senses Flavor 1977;2:375–90.
[27] Royet JP, Zald D, Versace R, et al. Emotional responses to pleasant
and unpleasant olfactory, visual and auditory stimuli: a PET study. J
Neurosci 2000;20:7752–9.
[28] Zald DH, Pardo JV. Functional neuroimaging of the olfactory
system in humans. Int J Psychophysiol 2003;36:165–81.