- Email: [email protected]
Olfactory event-related potentials in patients with brain tumors
Clinical Neurophysiology 112 (2001) 1523±1530
www.elsevier.com/locate/clinph
Olfactory event-related potentials in patients with brain tumors Christine Daniels a,*, Birgit Gottwald a, Bettina M. Pause b, Bernfried Sojka b, Hubertus M. Mehdorn a, Roman Ferstl b a
Department of Neurosurgery, University of Kiel, Kiel, Germany b Department of Psychology, University of Kiel, Kiel, Germany Accepted 20 April 2001
Abstract Objective: The aim of the study was to determine how odor processing is altered in patients with unilateral supratentorial brain tumors. Methods: Olfactory event-related potentials (OERPs) were evaluated in 10 patients with unilateral brain tumors of the frontal or temporal lobe in response to linalool and allylcaproate. Both odors were presented monorhinally by a constant-¯ow olfactometer. In addition, 20 healthy subjects were examined. While snif®ng, the subjects were asked to discriminate the two odors. EEG was recorded from 7 electrode positions (Fz, Cz, Pz, F3/4, P3/4). Amplitudes and latencies of 3 peaks (N1, P2, P3) were measured. To control for effects of modality-nonspeci®c alterations on the olfactory components acoustic event-related potentials (AERPs) were registered by use of an oddball paradigm. Results: Patients with right-sided lesions showed distinct de®cits in the discrimination task after stimulation of the right and left nostril. In contrast, patients with left-sided lesions only had an attenuation of correct reactions after left-sided stimulation. In the OERPs, patients with right-sided lesions showed P2- and P3-components with decreased amplitudes at parietal electrode positions. These alterations appeared after ipsi- and contralateral stimulation. Patients with left-sided lesions showed a signi®cant effect of the side of stimulation. Their OERPamplitudes were decreased after left-sided stimulation but not after right-sided stimulation. After right-sided olfactory stimulation a correlation between the olfactory and the acoustic ERP was seen in patients with right-sided lesions. Conclusions: Olfactory performance of the participating patients was markedly reduced. Patients with right-sided lesions showed bilateral impairment, which would support the importance of the right hemisphere in olfaction. The alteration of the topographic distribution of P2and P3-amplitudes in patients with right-sided lesions might re¯ect an impairment of early and late olfactory processing steps. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Brain tumor; Odor perception; Olfactory event-related potentials; P2; P3
1. Introduction Dysfunctions of the sense of smell have been detected in patients with various neurological diseases. In patients with Parkinsonian syndromes, the investigation of the olfactory performance can be helpful to distinguish idiopathic Parkinson's disease, which in most cases is accompanied by a markedly reduced olfaction from atypical Parkinsonism (Wenning et al., 1995). In patients with Alzheimer's, histopathological ®ndings as well as functional olfactory de®cits indicate that structural lesions of the olfactory cortex appear in an early stage of the disease. Evaluation of the olfactory system might be supportive in determining the diagnosis before severe cognitive de®cits are present (Solomon, * Corresponding author. Klinik fuÈr Neurochirurgie der UniversitaÈt Kiel, Weimarer Straûe 8, 24106 Kiel, Germany. Tel.: 149-431-5974924; fax: 149-431-5974884. E-mail address: [email protected] (C. Daniels).
1994). Also, a number of patients with brain lesions caused by trauma, vascular disease or neoplasms, just like patients after intracranial operations, suffer from an impairment of olfactory performance (Eslinger et al., 1982). In patients who have partial or incomplete loss of olfactory function, the de®cit may go completely unnoticed and may only become apparent when elaborate tests are used (Yousem et al., 1996). The olfactory event-related potentials (OERPs) have become established as an objective method in the investigation of (sub-) cortical odor processing. Due to its high time resolution, the OERP-technique allows the discrimination of single steps of odor processing, which are supposed to take place in certain brain regions and might represent different olfactory functions. Kobal and Hummel (1998) demonstrated the clinical signi®cance of OERPs in the assessment of anosmia. After olfactory stimulation with vanillin or hydrogen sul®de, no OERP-components were discernible in patients
1388-2457/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 1388-245 7(01)00590-9
CLINPH 2000814
1524
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
who suffered from anosmia for different reasons. Similar results were found in patients with congenital anosmia after stimulation with amyl acetate (Cui and Evans, 1997). However, Krauel et al. (1998) showed that experimental task demands and psychological state of the subject must be taken into account when evaluating and interpreting changes in the amplitude and latency of the OERP-components. Thus, a decline in the OERP-amplitudes can be caused by olfactory and by attentional disorders and careful evaluation is required to discriminate these effects. Pause et al. (1996) differentiated the ®rst two components, N1 and P2, which are mainly determined by stimulus characteristics (exogenous), from the P3-component, which is more dependent on endogenous processes. Since a correlation between odor detection threshold and P2-amplitude (Murphy et al., 1994) as well as a correlation between odor identi®cation and P2-amplitude (Hummel et al., 1998) and P2-latency (Evans et al., 1995) have been reported, an intimate relationship between the P2-component and exogenous olfactory processing may be suggested. In contrast, the P3-component varies depending on the subjective stimulus signi®cance and the stimulus probability (Pause et al., 1996), and might be modulated by the emotional valence of odors (Pause and Krauel, 2000). Geisler et al. (1999) observed an association between the P3-amplitudes/-latencies and subjects' performance on speci®c neuropsychological tasks. Due to its cognitive characteristics, the P3 of the OERP is suggested to re¯ect the features of the target associated P3 in other modalities. Although progress in neuronal imaging has allowed the study of functional topography of the CNS, brain regions activated in relation to speci®c olfactory processing steps have not been identi®ed precisely (Zald and Pardo, 2000). A positron emission tomography (PET) study on cerebral blood ¯ow during passive smelling showed a bilateral activation in the anterior temporal cortex and in the right orbitofrontal region (Zatorre et al., 1992). Medial temporal and orbitofrontal regions also have been detected by functional magnetic resonance imaging (fMRI), though the results of the different studies vary with regard to exact topographical data and hemispheric differences (Koizuka et al., 1994; Levy et al., 1997; Sobel et al., 1997; Yousem et al., 1997). Recently, magnetic source imaging has been used to determine those brain areas generating the components of the OERP (Kettenmann et al., 1997). Corresponding to the P2-component, a magnetically de®ned equivalent current dipole with localization in the superior temporal sulcus was obtained. This result suggested that apart from the known primary olfactory areas, neocortical areas are speci®cally involved in the processing of the olfactory information. Hummel et al. (1995) obtained OERPs in patients suffering from temporal lobe epilepsia. These patients showed well-de®ned brain lesions in the focus areas that had been localized and veri®ed by combined analysis of EEG and various neuroradiological imaging techniques. OERP-laten-
cies in patients with right-sided and left-sided epileptic focus were prolonged when the nostril ipsilateral to the epileptic focus was stimulated. Additionally, patients with right-sided temporal epileptic focus, but not patients with left-sided focus, showed a parieto-frontal shift of maximum P2-amplitudes in comparison to healthy subjects. Subjective testing of olfactory performance by means of an odor identi®cation test did not show differences between the two groups of patients. The aim of the present study was to de®ne the effects of unilateral supratentorial brain tumors on the OERP. Just like patients with epileptic foci, they suffered from focal CNSlesions of the CNS with the advantage that preoperatively assumed localization and size of the tumors could be veri®ed during operations. Based on evidence cited above, we hypothesized that patients with brain tumors show de®cits in olfactory tasks and that this impairment alters the P2- and P3-component of the OERP. Therefore, analysis of OERPdata will focus on the P2- and the P3-component. 2. Methods 2.1. Subjects Ten patients with unilateral brain tumors of the frontal or temporal lobe between 18 and 54 years of age (mean age 35 years; 5 men, 5 women) participated. They were recruited from the Department of Neurosurgery, University of Kiel, and had been screened for neuropsychological de®cits. Patients with severe de®cits of attention, concentration or general cognitive performance were excluded from the study. Three patients (P 3, 4, 8) had reported olfactory sensations associated to partial epileptic seizures, one patient (P 7) had recognized a loss of olfactory sensitivity after an earlier operation. Table 1 summarizes the patients' diagnoses and status (pre- or post-OP). Diagnoses are based on CT-/ MRIscans, operation reports, and histopathology. As a neuroradiological identi®cation of single olfactory structures from the CT-/MRI-scans with suf®cient reliability was not possible, the localization of the lesions was limited to the de®nition of the affected lobe and the side of hemisphere. In addition, 20 healthy subjects were examined (mean age 35 years, range 17±56 years; 10 men, 10 women). All of them had reported a normal sense of smell and their olfactory performance was veri®ed by a brief olfactory screening test (triple-forced-choice paradigm; target: linalool in a dilution of 1:50 in diethylphtalate). All subjects had no history of acute or chronic nasal infection or other diseases of the respiratory system and provided written informed consent. The study received prior approval by the ethics commission of the Medical School of the University of Kiel. 2.2. Odors and olfactometer Allylcaproate (97%, Morecombe, England) and linalool
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
1525
Table 1 Summary of investigated patients Patients
Age/gender
Diagnosis
Localization
Status
P P P P P P P P P P
26 years/male 37 years/male 28 years/male 40 years/female 18 years/male 31 years/female 36 years/female 37 years/female 42 years/male 54 years/female
Chondrosarcoma G I Astrocytoma WHO III Oligoastrocytoma WHO II Oligoastrocytoma WHO III Astrocytoma WHO II Astrocytoma WHO III Adenoma of the pituitary gland Astrocytoma WHO III Astrocytoma WHO II Glioblastoma multiform WHO IV
Left ala major ossis sphenoidalis Right parietal lobe Left frontal lobe Right medial temporal lobe Left medial temporal lobe Left frontal and temporal lobe Left sinus cavernosus Right medial temporal lobe Right basal temporal lobe Right temporal lobe
Pre-OP (status post partial exstirpation) Pre-OP Post two OPs and irradiation (60 Gy) Post-OP (10 days) Pre-OP Pre-OP Pre-OP Pre-OP Post-OP (6 days) Pre-OP
1 2 3 4 5 6 7 8 9 10
(97%, Merck-Schuchardt, Germany) were presented at a single concentration (dilution 1:50 with diethylphtalate). Both chemical stimulants have been described as odorants, which do not signi®cantly activate trigeminal receptors (Leplow et al., 1993; Doty et al., 1978). Their delivery was accomplished via a constant ¯ow olfactometer (Burghart Company, Germany). The method and design of the olfactometer have been described in detail by Kobal (1981). Stimuli were presented asynchronously to breathing within an air stream of constant ¯ow rate (100 ml/s), humidity (80%), and temperature (36.58C). Due to this stimulus presentation device, mechano- or thermoreceptors in the nasal mucosa were not activated signi®cantly. 2.3. Procedure After attaching the electrodes on their scalps, the subjects were comfortably seated in an air-conditioned room, which was monitored via a video camera. All the subjects were trained to perform a special breathing technique (`velopharyngeal closure', Kobal and Hummel, 1991), which was reported to improve the internal validity of the OERPrecording in comparison to spontaneous mouth breathing (Pause et al., 1999). At the beginning of the experiment, samples of the odorants were given via the olfactometer. The subjects learned to discriminate the two odors, coded as `odor 1' (linalool) and `odor 2' (allylcaproate), and to respond to each of them with a speci®c ®nger movement (odor 1: lifting the index ®nger of the right hand; odor 2: lifting the middle ®nger of the right hand). To avoid the possibility that the OERP could be overlapped by potentials related to the ®nger movement, the participants were asked to show the response not until an acoustic start signal was presented (latency 2±3 s after olfactory stimulus onset). In case of no sensation noted, subjects were asked to show no reaction. Odor samples were applied monorhinally, odor 1 always to the right nostril, odor 2 always to the left nostril. The test session lasted approximately 40 min and was subdivided into 4 sets of 20 stimuli each with 5 min rest periods between the sets. OERPs were obtained in response to randomized stimulation with allylcaproate and linalool. Each of the two stimuli (stimulus duration: 200 ms) was
presented 40 times (interstimulus interval: 15 s). In this session, both odors were presented to the right and to the left nostril in randomized order. The subjects' reactions were registered by a light sensor and the numbers of correct, false, and missing reactions were submitted to statistical analyses. White noise of 74 dB A was presented by headphones to mask sounds produced by the switching clicks of the olfactometer (Pause, personal communication). Subjects were asked to minimize head and eye movements. 2.4. OERPs The EEG (bandpass 0.016±30 Hz) was recorded unipolarly from 3 midline positions (Fz, Cz, Pz) and 4 lateral positions (F3/4, P3/4) of the 10/20-system referenced to link mastoids. To control for artifacts from eye movements, an electroocculogram (EOG) was registered from 5 additional electrode positions (Elbert et al., 1985). Single trials were screened for artifacts by off-line analysis and corrected from EOG effects by means of a multivariate regression model (Elbert et al., 1985). The mean of the 1000 ms prestimulus EEG recording served as a baseline for amplitude measurements. Averages were made for trials of the same side of stimulation. In the average OERPs, 3 peaks (N1, P2, P3) were determined in accordance to the characteristics of components described in Pause et al. (1996). Fig. 1 illustrates an example of an OERP of one
Fig. 1. OERP of one healthy subject after right-sided stimulation with allylcaproate n 20. Electrode position Pz, N1-/P2-/P3-components are labeled.
1526
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
healthy subject after right-sided stimulation with allylcaproate n 20 to demonstrate labeling of OERP-components. To avoid distortions of amplitude values by the arti®cially de®ned baseline, amplitudes of the P2- and P3-components were measured as peak-to-peak amplitudes N1±P2 and N1± P3 and these data were submitted to statistical tests. Additionally, base-to-peak amplitudes of the N1-component were submitted to similar tests in order to exclude that the N1±P2- and N1±P3-effects were in¯uenced by variations of N1.
calculated using Students t tests. Additionally, correlations between patients' OERP- and AERP-parameters at electrode position Cz were computed. Considering the different diagnoses of the patients, data of single patients were compared to data of control subjects separately. Therefore 95%-con®dence intervals were calculated for mean values of the control group and it was checked whether the single patients' corresponding values were within or outside these con®dence intervals.
2.5. AERPs
3. Results
To control for effects of modality-non-speci®c impairment on the olfactory components, AERPs were registered by use of a classical oddball paradigm. Subsequent to the olfactory stimulation, 120 acoustic stimuli (3 sets of 40 sine wave sounds, rests of 42 s between sets; stimulus duration 55 ms, interstimulus interval 4 s) consisting of 24 sounds of high frequency (target, 1200 Hz) and 96 sounds of low frequency (500 Hz) were presented binaurally by headphones. Patients were asked to count sounds of high frequency. EEG recording was performed under conditions similar to OERP recording. Single trials were screened for artifacts and corrected for EOG effects. Target trials were averaged and 3 peaks (N1, P2, P3) were labeled within the averaged waveform. Peak-to-peak amplitudes N1±P2 and N1±P3 were measured and submitted to statistical analyses. Base-to-peak amplitudes of N1 were used for control calculations analogous to the OERP-data.
3.1. Discrimination task
2.6. Statistical analyses For the comparison of OERP-parameters of patients and control subjects an analysis of variances (ANOVA) for repeated measurements including the within-subjects factors `side of stimulation (right±left nostril)', `electrode position in sagittal line (frontal±parietal)', and `electrode position in horizontal line (F/P3±F/Pz±F/P4)' was calculated across both odors. Results of the discrimination task were calculated by an ANOVA for repeated measurements with the within-subjects factor `reaction (correct±false± missing)' separately for right-sided and left-sided stimulation. In a previous step, the group of patients had been divided for the side of their brain lesion and data had been submitted to corresponding ANOVAs with the between-subjects factor `side of brain lesion (right±left hemisphere)' to check homogeneity in this group P . 0:20. In case of non-homogeneity, patients with right-sided lesions and patients with leftsided lesions were separately compared to the control group. Patients' AERP-data (averaged target trials) were submitted to an ANOVA for repeated measurements including the within-subjects factors `electrode position in sagittal line' and `electrode position in horizontal line' and the between-subjects factor `side of brain lesion'. Single comparisons in connection with all ANOVAs were
After right-sided stimulation, patients with right-sided lesions and patients with left-sided lesions had to be evaluated separately, because signi®cant effects for the interaction `side of brain lesion' £ `reaction' (F 2; 16 3:35, P , 0:20) had been found. In patients with right-sided lesions a signi®cant interaction `group' £ `reaction' (F 2; 26 14:46, P , 0:01) was revealed. In comparison to control subjects, patients with right-sided lesions showed a diminution of number of correct reactions (t 3:29, P , 0:05) and an increase in the number of missing reactions (t 2:94, P , 0:05). Patients with left-sided brain lesions did not differ from the control group. No signi®cant effect of the factor `side of brain lesion' was found after left-sided stimulation and patients were summarized to one group. Analyses exhibited a signi®cant interaction `group' £ `reaction' (F 2; 36 9:15, P , 0:01). Compared to healthy subjects, patients showed decreased numbers of correct reactions (t 3:25, P , 0:01) and increased numbers of missing reactions (t 23:49, P , 0:01). Fig. 2 illustrates the percentage of correct/missing reactions of healthy subjects, patients with right-sided lesions and patients with left-sided lesions after right-sided (100% 40 trials) and left-sided stimulation (100% 40 trials). For the incorrect reactions ( differences to 100%) no signi®cant effects have been found. 3.2. OERPs As signi®cant effects of the factor `side of brain lesion' were found (F 1; 8 2:33, P , 0:20 (N1P3-amplitude)), patients with right-sided and left-sided lesions had to be evaluated separately. In patients with right-sided lesions, no general diminution of mean peak amplitudes was observed, although there were found distinct deviations of amplitude values at certain electrode sites. N1P3-amplitudes showed a signi®cant interaction `group' £ `electrode position in sagittal line' (F 1; 23 9:09, P , 0:01). In contrast to the control group, where N1P3-amplitudes increased from frontal to parietal electrode positions, a parieto-frontal shift of maximum amplitudes with the highest N1P3-amplitudes at fron-
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
1527
3.3. AERPs
Fig. 2. Correct/missing reactions of healthy subjects, patients with rightsided lesions and patients with left-sided lesions after right-sided stimulation/left-sided stimulation.
tal positions was seen in patients with right-sided lesions. This effect was largest at right-sided electrode positions (interaction `group' £ `electrode position in sagittal line' £ ` electrode position in horizontal line' (F 2; 46 6:01, P , 0:01)). Also for the N1P2-amplitude a parieto-frontal shift of maximum amplitudes was found at right-sided electrode positions (interaction `group' £ `electrode position in sagittal line' £ `electrode position in horizontal line' (F 2; 46 7:41, P , 0:01)). No effect `side of stimulation' on the N1P2- or N1P3-amplitudes was found in patients with right-sided lesions. Patients with left-sided lesions did not show a parietofrontal shift of maximum P2- or P3-amplitudes. They exhibited large N1P3-amplitudes (main effect `group' (F 1; 23 3:31, P , 0:10)). This effect was most pronounced after right-sided stimulation and at right-sided electrode positions (signi®cant interactions `group' £ `side of stimulation' (F 1; 23 4:69, P , 0:05) and `group' £ ` side of stimulation' £ `electrode position in horizontal line' (F 2; 46 4:16, P , 0:05)). Fig. 3 illustrates the grandaveraged OERPs of healthy subjects, patients with rightsided lesions and patients with left-sided lesions at electrode position Pz. Values of single patients' amplitudes were compared to 95%-con®dence intervals of the control group. Single cases, in which maximum amplitudes were found at frontal electrode sites, were marked with `fz' (median sites: Fz . Pz), `f3' (left lateral sites: F3 . P3) and `f4' (right lateral sites: F4 . P4). Amplitudes were described as `decreased' if values of at least 4 electrode positions were below the corresponding intervals. A summary of this analysis is presented in Table 2.
P3-amplitudes of patients with left-sided lesions showed signi®cant effects of the factors `electrode position in sagittal line' (F 1; 3 78:19, P , 0:01) and `electrode positions in horizontal line' (F 2; 6 13:51, P , 0:01), indicating that P3-amplitudes were generally largest at parietal and median electrode sites. In contrast, no signi®cant effects of the factors `electrode position in sagittal line' or `electrode positions in horizontal line' were found in patients with right-sided lesions. These patients did not produce the characteristic pattern of topographical distribution of the P3-amplitude that had been described for AERPs of healthy subjects in numerous studies. In addition, a signi®cant correlation between P3-amplitudes of the AERP and the OERP of patients with right-sided lesions was found after right-sided olfactory stimulation (r 0:93, P 0:02). In contrast, no correlation was discernible for the N1- and P2-amplitudes of the OERP and the AERP. Fig. 4 shows grand average AERPs at electrode positions Fz, Cz, and Pz in patients with right- and left-sided lesions.
Fig. 3. (a) Grand average OERPs of healthy subjects, patients with rightsided lesions and patients with left-sided lesions at electrode position Pz ± right-sided stimulation. (b) Grand average OERPs of healthy subjects, patients with right-sided lesions and patients with left-sided lesions at electrode position Pz ± left-sided stimulation.
1528
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
Table 2 Distribution of maximum amplitudes and amplitude values of single patients a Patients with right-sided lesions Stimulated nostril
Patients with left-sided lesions
N1P2-amplitudes Right
N1P3-amplitudes Left
Right
(A) Distribution of maximum amplitude of single patients (P 1±P 10) P2 fz, f3, f4 f3, f4 P4 f4 P8 fz, f4 fz, f3, f4 P9 f4 fz, f3, f4 P 10 f4 f4
Left
f4 f4
P P P P P
1 3 5 6 7
(B) Signi®cant changes of amplitudes: comparisons of patients (P 1±P 10) with healthy controls P2 # # P1 P4 # # # # P3 P8 P5 P9 # P6 P 10 # # # # P7
N1P2-amplitudes
N1P3-amplitudes
Right
Left
Right
f4
f4 f4 fz, f3, f4 f3
Left
fz, f3, f4 f3
# # #
#
a
fz, amplitude values at Fz . amplitude values at Pz; f3, amplitude values at F3 . amplitude values at P3; f4, amplitude values at F4 . amplitude values at P4. # , decreased
4. Discussion The present investigation indicates marked de®cits in olfactory performance in patients with supratentorial brain tumors. Results of the discrimination task showed decreased numbers of correct reactions in both patients with rightsided and patients with left-sided lesions. However, in¯uence of the side of stimulation differed in both groups: whereas patients with left-sided lesions only had an attenua-
tion of correct reactions after left-sided, i.e. ipsilateral, stimulation, patients with right-sided lesions showed reduced numbers of correct reactions after right- and leftsided stimulation. This ®nding is in accordance with the results of numerous studies (e.g. Zatorre and Jones-Gotman, 1991) that reported greater olfactory impairment in patients with lesions of the right hemisphere in comparison to patients with analogous lesions of the left hemisphere. Also, present OERP-data of patients with right-sided lesions
Fig. 4. Grand average AERPs at electrode positions Fz, Cz, and Pz in patients with right-sided lesions/left-sided lesions.
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
indicate that structural lesions of the right hemisphere may alter olfactory processing steps after right-sided and leftsided stimulation. In these patients, an alteration of topographical distribution of maximum amplitudes could be observed independent of the side of stimulation. Hummel et al. (1995) found comparable results in patients with rightsided temporal lobe epilepsy. Additionally, they reported latency differences (prolonged latencies after ipsilateral stimulation) between patients with right- or left-sided epileptic foci to be more pronounced after stimulation of the right nostril, which also might be related to more severe olfactory impairment after lesions of the right hemisphere. It is remarkable that only one of the patients in the present study had recognized an impairment of olfactory functions in advance. Doty et al. (1997) reported that many patients with olfactory loss secondary to head trauma were inaccurate in judging the state of their chemosensory functioning. Also, patients with unilateral anosmia after neurosurgical operations via the fronto-temporal approach most often did not recognize this complication (Eriksen et al., 1990). These disparities between subjective impression and objective test results seem to be characteristic for disturbances of the olfactory system and also have been described in patients with chronic neurological disease, e.g. Parkinson's disease (Doty et al., 1988). Considering special electrode positions, the patients' OERP-data showed distinct differences compared to data of the control group. In contrast, an evaluation of mean amplitude values of the 7 recording sites exhibited no significant effects. This ®nding emphasizes the importance of EEG recording from numerous electrode positions and a position-related analysis of OERP-data. In patients with right-sided lesions, decreased P2- and P3amplitudes could be found at parietal electrode positions. At the frontal recording sites, amplitudes did not show differences in comparison to the control group. In contrast to healthy subjects, where amplitudes increased from frontal to parietal sites, maximum amplitude values of patients with right-sided lesions were measured at frontal recording sites. These ®ndings compare to the observations that had been made in healthy elderly subjects as well as in patients suffering from right-sided temporal lobe epilepsy. Several authors reported age-related effects of topographical distribution of P2-amplitudes: OERP-data of the elderly (mean 66 years) in response to amyl acetate showed differences from data of young adults (mean 27 years) in P2-amplitudes at recording positions Cz and Pz, but not at the frontal site Fz (Murphy et al.,1994). In young subjects, amplitude increased from Fz to Cz to Pz; elderly subjects showed no such increase. Similar observations have been made for the P3-component (Murphy et al., 1998; Geisler et al., 1999). Hummel et al. (1998) investigated subjects of 3 age ranges (15±34, 35±54, and 55±74 years) and reported that the fronto-parietal topographical differences between the OERP-amplitudes in response to vanillin and hydrogen sul®de leveled off with increasing age. A comparable observation was made in
1529
patients with right-sided temporal lobe epilepsy (Hummel et al., 1995). In contrast to patients with a left-sided focus, who showed normal distribution of maximum P2-amplitudes, these patients elicited largest P2-amplitudes at Cz. This shift of topographical distribution of amplitudes might indicate an alteration of direction or location of the dipoles producing the OERP. A change of dipole characteristics could be expected, when brain regions that are normally involved in olfactory processing steps are destroyed by brain tumors as well as by epileptic seizures or neurodegenerative disease. The question remains whether the observed electrophysiological ®ndings re¯ect an impairment of olfactory processing or just represent an epiphenomenon of modality-non-speci®c changes of the CNS, which could be expected as a result of the brain tumor, the edema, and the destruction of neuronal structures. Our ®ndings concerning patients' impairment in the olfactory discrimination task on one hand and alteration of the P2-amplitude on the other hand would support a modality-speci®c characteristic of this OERP-component as suggested by several authors (Murphy et al., 1994; Evans et al., 1995; Hummel et al., 1998). In contrast, the observed correlation between the P3-amplitudes of OERP and AERP in patients with right-sided brain lesions after ipsilateral stimulation indicates that this component might re¯ect modality-non-speci®c processing steps in the right hemisphere. The hypothesis of a modalitynon-speci®c characteristic of the P3 was already made by Pause et al. (1996) and had been supported by following studies (Murphy et al., 1998; Geisler et al., 1999). Nevertheless, as the processing steps underlying the P3-component of the OERP are associated with olfactory stimuli, they might give indirect information about an early sensory perception of an odor, which is later processed within associative cortex areas. Based on this assumption, the alteration of topographical distribution of the P3-amplitudes in patients with right-sided brain lesions, which was observed after ipsi- and contralateral stimulation, also would re¯ect the bilateral olfactory impairment of these patients. In conclusion, investigation of OERPs in patients with brain lesions can give information on the functionality of the olfactory system. The diagnostic value of the OERPs lies in the possibility of detecting unnoticed functional olfactory de®cits, screening neurological or psychiatric patients. Nevertheless, data of single patients have to be interpreted carefully and deviant ®ndings should be seen as an indication for further clinical olfactory tests. Acknowledgements We thank K. Krauel, T. Ehmke, C. MuÈller and K. Rogalski for their technical assistance. Preliminary results from this study were presented at the symposium `Chemosensory Bioresponses in Man II', held at Erlangen, Germany, November/December 1999.
1530
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
References Cui L, Evans WJ. Olfactory event-related potentials to amyl acetate in congenital anosmia. Electroenceph clin Neurophysiol 1997;102:303± 306. Doty RL, Brugger WE, Jurs PC, Ornolorff MA, Snyder PJ, Lowry LD. Intranasal trigeminal stimulation from odor volatiles. Psychometric responses from anosmics and normal human. Physiol Behav 1978;20:175±185. Doty RL, Deems DA, Stellar S. Olfactory dysfunction in Parkinson's desease. A general de®cit unrelated to neurologic signs, disease stage or disease duration. Neurology 1988;38:1237±1244. Doty RL, Yousem DM, Pham LT, Kreshak AA, Geckle R, Lee WW. Olfactory dysfunctions in patients with head trauma. Arch Neurol 1997;54:1131±1140. Elbert T, Lutzenberger W, Rockstroh B, Birnbaumer N. Removal of ocular artifacts from the EEG ± a biophysical approach. Electroenceph clin Neurophysiol 1985;60:455±463. Eriksen KD, Boge-Rasmussen T, Kruse-Larsen C. Anosmia following operation for cerebral aneurysms in the anterior circulation. J Neurosurg 1990;72:864±865. Eslinger PJ, Damasio AR, Van Hoesen GW. Olfactory dysfunction in man: anatomical and behavioral aspects. Brain Cogn 1982;1:259±285. Evans J, Cui L, Starr A. Olfactory event-related potentials in normal human subjects. Effects of age and gender. Electroenceph clin Neurophysiol 1995;95:293±301. Geisler MW, Morgan CD, Covington JW. Neuropsychological performance and cognitive event-related brain potentials in young and elderly adults. Clin Exp Neuropsychol 1999;211:108±126. Hummel T, Pauli E, SchuÈler P, Kettenmann B, Stefan H, Kobal G. Chemosensory event-related potentials in patients with temporal lobe epilepsy. Epilepsia 1995;361:79±85. Hummel T, Barz S, Pauli E, Kobal G. Chemosensory event-related potentials change with age. Electroenceph clin Neurophysiol 1998;108:208± 217. Kettenmann B, Hummel C, Stefan H, Kobal G. Multiple olfactory activity in the human neocortex identi®ed by magnetic source imaging. Chem Senses 1997;22:493±502. Kobal G. Elektrophysiologische Untersuchungen des menschlichen Geruchssinnes. 1st ed. Stuttgart: Thieme, 1981. Kobal G, Hummel T. Olfactory evoked potentials in humans. In: Getchell TV, Doty RL, Bartoshuk LM, Snow JB, editors. Smell and taste in health and disease, New York, NY: Raven Press, 1991. pp. 255±275. Kobal G, Hummel T. Olfactory and intranasal trigeminal event-related potentials in anosmic patients. Laryngoscope 1998;108:1033±1035. Koizuka I, Yano H, Nagahara M, Mochizuki R, Seo R, Shimada K, Kubo T,
Nogawa T. Functional imaging of the human olfactory cortex by magnetic resonance imaging. Otorhinolaryngology 1994;56:273±275. Krauel K, Pause BM, Sojka B, Schott P, Ferstl R. Attentional modulation of central odor processing. Chem Senses 1998;23:423±432. Leplow B, Pause BM, Behrens C, Sojka B, Ferstl R, Mehdorn HM. Assessment of olfactory and trigeminal perception of odors in patients with lesions of the olfactory system. Poster at the 21st meeting of the German Society of Psychophysiology and its Applications, WuÈrzburg, June 9± 12, 1993. Levy LM, Henkin RI, Lin CS, Martins D, Schellinger D. Functional MRI of human olfaction. J Comput Assist Tomogr 1997;21(6):849±856. Murphy C, Nordin S, de Wijk RA, Cain WS, Polich J. Olfactory-evoked potentials: assessment of young and elderly, and comparison to psychophysical threshold. Chem Senses 1994;19:47±56. Murphy C, Wetter S, Morgan CD, Ellison DW, Geisler MW. Age effects on central nervous system activity re¯ected in the olfactory event-related potential. Ann N Y Acad Sci 1998;855:598±607. Pause BM, Krauel K. Chemosensory event-related potentials (CSERP) as a key to the psychology of odors. Int J Psychophysiol 2000;36:105±122. Pause BM, Sojka B, Krauel K, Ferstl R. The nature of the late positive complex within the olfactory event-related potential OERP. Psychophysiology 1996;33:376±384. Pause BM, Krauel K, Sojka B, Ferstl R. Is odor processing related to oral breathing? Int J Psychophysiol 1999;32:251±260. Sobel N, Prabhakaran V, Desmond JE, Glover GH, Sullivan EV, Gabrieli JDE. A method for functional magnetic resonance imaging of olfaction. J Neurosci Methods 1997;78:115±123. Solomon GS. Anosmia in Alzheimer disease. Percept Motor Skills 1994;79:1249±1250. Wenning GK, Shephard B, Hawkes C, Petruckevitch A, Lees A, Quinn N. Olfactory function in atypical parkinsonian syndromes. Acta Neurol Scand 1995;91:247±250. Yousem DM, Geckle RJ, Bilker WB, McKeown DA, Doty RL. Posttraumatic olfactory dysfunction. MR and clinical evaluation. Am J Neuroradiol 1996;17:1171±1179. Yousem DM, Williams SCR, Howard RO, Andrew C, Simmons A, Allin M, Geckle R, Suskind D, Bullmore ET, Brammer MJ, Doty RL. Functional MR imaging during odor stimulation. Preliminary data. Radiology 1997;204:833±838. Zald DH, Pardo JV. Functional neuroimaging of the olfactory system in humans. Int J Psychophysiol 2000;36:165±181. Zatorre RJ, Jones-Gotman M. Human olfactory discrimination after unilateral frontal or temporal lobectomy. Brain 1991;114:71±84. Zatorre RJ, Jones-Gotman M, Evans AC, Meyer E. Functional localization and lateralization of human olfactory cortex. Nature 1992;360:339± 340.
www.elsevier.com/locate/clinph
Olfactory event-related potentials in patients with brain tumors Christine Daniels a,*, Birgit Gottwald a, Bettina M. Pause b, Bernfried Sojka b, Hubertus M. Mehdorn a, Roman Ferstl b a
Department of Neurosurgery, University of Kiel, Kiel, Germany b Department of Psychology, University of Kiel, Kiel, Germany Accepted 20 April 2001
Abstract Objective: The aim of the study was to determine how odor processing is altered in patients with unilateral supratentorial brain tumors. Methods: Olfactory event-related potentials (OERPs) were evaluated in 10 patients with unilateral brain tumors of the frontal or temporal lobe in response to linalool and allylcaproate. Both odors were presented monorhinally by a constant-¯ow olfactometer. In addition, 20 healthy subjects were examined. While snif®ng, the subjects were asked to discriminate the two odors. EEG was recorded from 7 electrode positions (Fz, Cz, Pz, F3/4, P3/4). Amplitudes and latencies of 3 peaks (N1, P2, P3) were measured. To control for effects of modality-nonspeci®c alterations on the olfactory components acoustic event-related potentials (AERPs) were registered by use of an oddball paradigm. Results: Patients with right-sided lesions showed distinct de®cits in the discrimination task after stimulation of the right and left nostril. In contrast, patients with left-sided lesions only had an attenuation of correct reactions after left-sided stimulation. In the OERPs, patients with right-sided lesions showed P2- and P3-components with decreased amplitudes at parietal electrode positions. These alterations appeared after ipsi- and contralateral stimulation. Patients with left-sided lesions showed a signi®cant effect of the side of stimulation. Their OERPamplitudes were decreased after left-sided stimulation but not after right-sided stimulation. After right-sided olfactory stimulation a correlation between the olfactory and the acoustic ERP was seen in patients with right-sided lesions. Conclusions: Olfactory performance of the participating patients was markedly reduced. Patients with right-sided lesions showed bilateral impairment, which would support the importance of the right hemisphere in olfaction. The alteration of the topographic distribution of P2and P3-amplitudes in patients with right-sided lesions might re¯ect an impairment of early and late olfactory processing steps. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Brain tumor; Odor perception; Olfactory event-related potentials; P2; P3
1. Introduction Dysfunctions of the sense of smell have been detected in patients with various neurological diseases. In patients with Parkinsonian syndromes, the investigation of the olfactory performance can be helpful to distinguish idiopathic Parkinson's disease, which in most cases is accompanied by a markedly reduced olfaction from atypical Parkinsonism (Wenning et al., 1995). In patients with Alzheimer's, histopathological ®ndings as well as functional olfactory de®cits indicate that structural lesions of the olfactory cortex appear in an early stage of the disease. Evaluation of the olfactory system might be supportive in determining the diagnosis before severe cognitive de®cits are present (Solomon, * Corresponding author. Klinik fuÈr Neurochirurgie der UniversitaÈt Kiel, Weimarer Straûe 8, 24106 Kiel, Germany. Tel.: 149-431-5974924; fax: 149-431-5974884. E-mail address: [email protected] (C. Daniels).
1994). Also, a number of patients with brain lesions caused by trauma, vascular disease or neoplasms, just like patients after intracranial operations, suffer from an impairment of olfactory performance (Eslinger et al., 1982). In patients who have partial or incomplete loss of olfactory function, the de®cit may go completely unnoticed and may only become apparent when elaborate tests are used (Yousem et al., 1996). The olfactory event-related potentials (OERPs) have become established as an objective method in the investigation of (sub-) cortical odor processing. Due to its high time resolution, the OERP-technique allows the discrimination of single steps of odor processing, which are supposed to take place in certain brain regions and might represent different olfactory functions. Kobal and Hummel (1998) demonstrated the clinical signi®cance of OERPs in the assessment of anosmia. After olfactory stimulation with vanillin or hydrogen sul®de, no OERP-components were discernible in patients
1388-2457/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 1388-245 7(01)00590-9
CLINPH 2000814
1524
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
who suffered from anosmia for different reasons. Similar results were found in patients with congenital anosmia after stimulation with amyl acetate (Cui and Evans, 1997). However, Krauel et al. (1998) showed that experimental task demands and psychological state of the subject must be taken into account when evaluating and interpreting changes in the amplitude and latency of the OERP-components. Thus, a decline in the OERP-amplitudes can be caused by olfactory and by attentional disorders and careful evaluation is required to discriminate these effects. Pause et al. (1996) differentiated the ®rst two components, N1 and P2, which are mainly determined by stimulus characteristics (exogenous), from the P3-component, which is more dependent on endogenous processes. Since a correlation between odor detection threshold and P2-amplitude (Murphy et al., 1994) as well as a correlation between odor identi®cation and P2-amplitude (Hummel et al., 1998) and P2-latency (Evans et al., 1995) have been reported, an intimate relationship between the P2-component and exogenous olfactory processing may be suggested. In contrast, the P3-component varies depending on the subjective stimulus signi®cance and the stimulus probability (Pause et al., 1996), and might be modulated by the emotional valence of odors (Pause and Krauel, 2000). Geisler et al. (1999) observed an association between the P3-amplitudes/-latencies and subjects' performance on speci®c neuropsychological tasks. Due to its cognitive characteristics, the P3 of the OERP is suggested to re¯ect the features of the target associated P3 in other modalities. Although progress in neuronal imaging has allowed the study of functional topography of the CNS, brain regions activated in relation to speci®c olfactory processing steps have not been identi®ed precisely (Zald and Pardo, 2000). A positron emission tomography (PET) study on cerebral blood ¯ow during passive smelling showed a bilateral activation in the anterior temporal cortex and in the right orbitofrontal region (Zatorre et al., 1992). Medial temporal and orbitofrontal regions also have been detected by functional magnetic resonance imaging (fMRI), though the results of the different studies vary with regard to exact topographical data and hemispheric differences (Koizuka et al., 1994; Levy et al., 1997; Sobel et al., 1997; Yousem et al., 1997). Recently, magnetic source imaging has been used to determine those brain areas generating the components of the OERP (Kettenmann et al., 1997). Corresponding to the P2-component, a magnetically de®ned equivalent current dipole with localization in the superior temporal sulcus was obtained. This result suggested that apart from the known primary olfactory areas, neocortical areas are speci®cally involved in the processing of the olfactory information. Hummel et al. (1995) obtained OERPs in patients suffering from temporal lobe epilepsia. These patients showed well-de®ned brain lesions in the focus areas that had been localized and veri®ed by combined analysis of EEG and various neuroradiological imaging techniques. OERP-laten-
cies in patients with right-sided and left-sided epileptic focus were prolonged when the nostril ipsilateral to the epileptic focus was stimulated. Additionally, patients with right-sided temporal epileptic focus, but not patients with left-sided focus, showed a parieto-frontal shift of maximum P2-amplitudes in comparison to healthy subjects. Subjective testing of olfactory performance by means of an odor identi®cation test did not show differences between the two groups of patients. The aim of the present study was to de®ne the effects of unilateral supratentorial brain tumors on the OERP. Just like patients with epileptic foci, they suffered from focal CNSlesions of the CNS with the advantage that preoperatively assumed localization and size of the tumors could be veri®ed during operations. Based on evidence cited above, we hypothesized that patients with brain tumors show de®cits in olfactory tasks and that this impairment alters the P2- and P3-component of the OERP. Therefore, analysis of OERPdata will focus on the P2- and the P3-component. 2. Methods 2.1. Subjects Ten patients with unilateral brain tumors of the frontal or temporal lobe between 18 and 54 years of age (mean age 35 years; 5 men, 5 women) participated. They were recruited from the Department of Neurosurgery, University of Kiel, and had been screened for neuropsychological de®cits. Patients with severe de®cits of attention, concentration or general cognitive performance were excluded from the study. Three patients (P 3, 4, 8) had reported olfactory sensations associated to partial epileptic seizures, one patient (P 7) had recognized a loss of olfactory sensitivity after an earlier operation. Table 1 summarizes the patients' diagnoses and status (pre- or post-OP). Diagnoses are based on CT-/ MRIscans, operation reports, and histopathology. As a neuroradiological identi®cation of single olfactory structures from the CT-/MRI-scans with suf®cient reliability was not possible, the localization of the lesions was limited to the de®nition of the affected lobe and the side of hemisphere. In addition, 20 healthy subjects were examined (mean age 35 years, range 17±56 years; 10 men, 10 women). All of them had reported a normal sense of smell and their olfactory performance was veri®ed by a brief olfactory screening test (triple-forced-choice paradigm; target: linalool in a dilution of 1:50 in diethylphtalate). All subjects had no history of acute or chronic nasal infection or other diseases of the respiratory system and provided written informed consent. The study received prior approval by the ethics commission of the Medical School of the University of Kiel. 2.2. Odors and olfactometer Allylcaproate (97%, Morecombe, England) and linalool
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
1525
Table 1 Summary of investigated patients Patients
Age/gender
Diagnosis
Localization
Status
P P P P P P P P P P
26 years/male 37 years/male 28 years/male 40 years/female 18 years/male 31 years/female 36 years/female 37 years/female 42 years/male 54 years/female
Chondrosarcoma G I Astrocytoma WHO III Oligoastrocytoma WHO II Oligoastrocytoma WHO III Astrocytoma WHO II Astrocytoma WHO III Adenoma of the pituitary gland Astrocytoma WHO III Astrocytoma WHO II Glioblastoma multiform WHO IV
Left ala major ossis sphenoidalis Right parietal lobe Left frontal lobe Right medial temporal lobe Left medial temporal lobe Left frontal and temporal lobe Left sinus cavernosus Right medial temporal lobe Right basal temporal lobe Right temporal lobe
Pre-OP (status post partial exstirpation) Pre-OP Post two OPs and irradiation (60 Gy) Post-OP (10 days) Pre-OP Pre-OP Pre-OP Pre-OP Post-OP (6 days) Pre-OP
1 2 3 4 5 6 7 8 9 10
(97%, Merck-Schuchardt, Germany) were presented at a single concentration (dilution 1:50 with diethylphtalate). Both chemical stimulants have been described as odorants, which do not signi®cantly activate trigeminal receptors (Leplow et al., 1993; Doty et al., 1978). Their delivery was accomplished via a constant ¯ow olfactometer (Burghart Company, Germany). The method and design of the olfactometer have been described in detail by Kobal (1981). Stimuli were presented asynchronously to breathing within an air stream of constant ¯ow rate (100 ml/s), humidity (80%), and temperature (36.58C). Due to this stimulus presentation device, mechano- or thermoreceptors in the nasal mucosa were not activated signi®cantly. 2.3. Procedure After attaching the electrodes on their scalps, the subjects were comfortably seated in an air-conditioned room, which was monitored via a video camera. All the subjects were trained to perform a special breathing technique (`velopharyngeal closure', Kobal and Hummel, 1991), which was reported to improve the internal validity of the OERPrecording in comparison to spontaneous mouth breathing (Pause et al., 1999). At the beginning of the experiment, samples of the odorants were given via the olfactometer. The subjects learned to discriminate the two odors, coded as `odor 1' (linalool) and `odor 2' (allylcaproate), and to respond to each of them with a speci®c ®nger movement (odor 1: lifting the index ®nger of the right hand; odor 2: lifting the middle ®nger of the right hand). To avoid the possibility that the OERP could be overlapped by potentials related to the ®nger movement, the participants were asked to show the response not until an acoustic start signal was presented (latency 2±3 s after olfactory stimulus onset). In case of no sensation noted, subjects were asked to show no reaction. Odor samples were applied monorhinally, odor 1 always to the right nostril, odor 2 always to the left nostril. The test session lasted approximately 40 min and was subdivided into 4 sets of 20 stimuli each with 5 min rest periods between the sets. OERPs were obtained in response to randomized stimulation with allylcaproate and linalool. Each of the two stimuli (stimulus duration: 200 ms) was
presented 40 times (interstimulus interval: 15 s). In this session, both odors were presented to the right and to the left nostril in randomized order. The subjects' reactions were registered by a light sensor and the numbers of correct, false, and missing reactions were submitted to statistical analyses. White noise of 74 dB A was presented by headphones to mask sounds produced by the switching clicks of the olfactometer (Pause, personal communication). Subjects were asked to minimize head and eye movements. 2.4. OERPs The EEG (bandpass 0.016±30 Hz) was recorded unipolarly from 3 midline positions (Fz, Cz, Pz) and 4 lateral positions (F3/4, P3/4) of the 10/20-system referenced to link mastoids. To control for artifacts from eye movements, an electroocculogram (EOG) was registered from 5 additional electrode positions (Elbert et al., 1985). Single trials were screened for artifacts by off-line analysis and corrected from EOG effects by means of a multivariate regression model (Elbert et al., 1985). The mean of the 1000 ms prestimulus EEG recording served as a baseline for amplitude measurements. Averages were made for trials of the same side of stimulation. In the average OERPs, 3 peaks (N1, P2, P3) were determined in accordance to the characteristics of components described in Pause et al. (1996). Fig. 1 illustrates an example of an OERP of one
Fig. 1. OERP of one healthy subject after right-sided stimulation with allylcaproate n 20. Electrode position Pz, N1-/P2-/P3-components are labeled.
1526
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
healthy subject after right-sided stimulation with allylcaproate n 20 to demonstrate labeling of OERP-components. To avoid distortions of amplitude values by the arti®cially de®ned baseline, amplitudes of the P2- and P3-components were measured as peak-to-peak amplitudes N1±P2 and N1± P3 and these data were submitted to statistical tests. Additionally, base-to-peak amplitudes of the N1-component were submitted to similar tests in order to exclude that the N1±P2- and N1±P3-effects were in¯uenced by variations of N1.
calculated using Students t tests. Additionally, correlations between patients' OERP- and AERP-parameters at electrode position Cz were computed. Considering the different diagnoses of the patients, data of single patients were compared to data of control subjects separately. Therefore 95%-con®dence intervals were calculated for mean values of the control group and it was checked whether the single patients' corresponding values were within or outside these con®dence intervals.
2.5. AERPs
3. Results
To control for effects of modality-non-speci®c impairment on the olfactory components, AERPs were registered by use of a classical oddball paradigm. Subsequent to the olfactory stimulation, 120 acoustic stimuli (3 sets of 40 sine wave sounds, rests of 42 s between sets; stimulus duration 55 ms, interstimulus interval 4 s) consisting of 24 sounds of high frequency (target, 1200 Hz) and 96 sounds of low frequency (500 Hz) were presented binaurally by headphones. Patients were asked to count sounds of high frequency. EEG recording was performed under conditions similar to OERP recording. Single trials were screened for artifacts and corrected for EOG effects. Target trials were averaged and 3 peaks (N1, P2, P3) were labeled within the averaged waveform. Peak-to-peak amplitudes N1±P2 and N1±P3 were measured and submitted to statistical analyses. Base-to-peak amplitudes of N1 were used for control calculations analogous to the OERP-data.
3.1. Discrimination task
2.6. Statistical analyses For the comparison of OERP-parameters of patients and control subjects an analysis of variances (ANOVA) for repeated measurements including the within-subjects factors `side of stimulation (right±left nostril)', `electrode position in sagittal line (frontal±parietal)', and `electrode position in horizontal line (F/P3±F/Pz±F/P4)' was calculated across both odors. Results of the discrimination task were calculated by an ANOVA for repeated measurements with the within-subjects factor `reaction (correct±false± missing)' separately for right-sided and left-sided stimulation. In a previous step, the group of patients had been divided for the side of their brain lesion and data had been submitted to corresponding ANOVAs with the between-subjects factor `side of brain lesion (right±left hemisphere)' to check homogeneity in this group P . 0:20. In case of non-homogeneity, patients with right-sided lesions and patients with leftsided lesions were separately compared to the control group. Patients' AERP-data (averaged target trials) were submitted to an ANOVA for repeated measurements including the within-subjects factors `electrode position in sagittal line' and `electrode position in horizontal line' and the between-subjects factor `side of brain lesion'. Single comparisons in connection with all ANOVAs were
After right-sided stimulation, patients with right-sided lesions and patients with left-sided lesions had to be evaluated separately, because signi®cant effects for the interaction `side of brain lesion' £ `reaction' (F 2; 16 3:35, P , 0:20) had been found. In patients with right-sided lesions a signi®cant interaction `group' £ `reaction' (F 2; 26 14:46, P , 0:01) was revealed. In comparison to control subjects, patients with right-sided lesions showed a diminution of number of correct reactions (t 3:29, P , 0:05) and an increase in the number of missing reactions (t 2:94, P , 0:05). Patients with left-sided brain lesions did not differ from the control group. No signi®cant effect of the factor `side of brain lesion' was found after left-sided stimulation and patients were summarized to one group. Analyses exhibited a signi®cant interaction `group' £ `reaction' (F 2; 36 9:15, P , 0:01). Compared to healthy subjects, patients showed decreased numbers of correct reactions (t 3:25, P , 0:01) and increased numbers of missing reactions (t 23:49, P , 0:01). Fig. 2 illustrates the percentage of correct/missing reactions of healthy subjects, patients with right-sided lesions and patients with left-sided lesions after right-sided (100% 40 trials) and left-sided stimulation (100% 40 trials). For the incorrect reactions ( differences to 100%) no signi®cant effects have been found. 3.2. OERPs As signi®cant effects of the factor `side of brain lesion' were found (F 1; 8 2:33, P , 0:20 (N1P3-amplitude)), patients with right-sided and left-sided lesions had to be evaluated separately. In patients with right-sided lesions, no general diminution of mean peak amplitudes was observed, although there were found distinct deviations of amplitude values at certain electrode sites. N1P3-amplitudes showed a signi®cant interaction `group' £ `electrode position in sagittal line' (F 1; 23 9:09, P , 0:01). In contrast to the control group, where N1P3-amplitudes increased from frontal to parietal electrode positions, a parieto-frontal shift of maximum amplitudes with the highest N1P3-amplitudes at fron-
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
1527
3.3. AERPs
Fig. 2. Correct/missing reactions of healthy subjects, patients with rightsided lesions and patients with left-sided lesions after right-sided stimulation/left-sided stimulation.
tal positions was seen in patients with right-sided lesions. This effect was largest at right-sided electrode positions (interaction `group' £ `electrode position in sagittal line' £ ` electrode position in horizontal line' (F 2; 46 6:01, P , 0:01)). Also for the N1P2-amplitude a parieto-frontal shift of maximum amplitudes was found at right-sided electrode positions (interaction `group' £ `electrode position in sagittal line' £ `electrode position in horizontal line' (F 2; 46 7:41, P , 0:01)). No effect `side of stimulation' on the N1P2- or N1P3-amplitudes was found in patients with right-sided lesions. Patients with left-sided lesions did not show a parietofrontal shift of maximum P2- or P3-amplitudes. They exhibited large N1P3-amplitudes (main effect `group' (F 1; 23 3:31, P , 0:10)). This effect was most pronounced after right-sided stimulation and at right-sided electrode positions (signi®cant interactions `group' £ `side of stimulation' (F 1; 23 4:69, P , 0:05) and `group' £ ` side of stimulation' £ `electrode position in horizontal line' (F 2; 46 4:16, P , 0:05)). Fig. 3 illustrates the grandaveraged OERPs of healthy subjects, patients with rightsided lesions and patients with left-sided lesions at electrode position Pz. Values of single patients' amplitudes were compared to 95%-con®dence intervals of the control group. Single cases, in which maximum amplitudes were found at frontal electrode sites, were marked with `fz' (median sites: Fz . Pz), `f3' (left lateral sites: F3 . P3) and `f4' (right lateral sites: F4 . P4). Amplitudes were described as `decreased' if values of at least 4 electrode positions were below the corresponding intervals. A summary of this analysis is presented in Table 2.
P3-amplitudes of patients with left-sided lesions showed signi®cant effects of the factors `electrode position in sagittal line' (F 1; 3 78:19, P , 0:01) and `electrode positions in horizontal line' (F 2; 6 13:51, P , 0:01), indicating that P3-amplitudes were generally largest at parietal and median electrode sites. In contrast, no signi®cant effects of the factors `electrode position in sagittal line' or `electrode positions in horizontal line' were found in patients with right-sided lesions. These patients did not produce the characteristic pattern of topographical distribution of the P3-amplitude that had been described for AERPs of healthy subjects in numerous studies. In addition, a signi®cant correlation between P3-amplitudes of the AERP and the OERP of patients with right-sided lesions was found after right-sided olfactory stimulation (r 0:93, P 0:02). In contrast, no correlation was discernible for the N1- and P2-amplitudes of the OERP and the AERP. Fig. 4 shows grand average AERPs at electrode positions Fz, Cz, and Pz in patients with right- and left-sided lesions.
Fig. 3. (a) Grand average OERPs of healthy subjects, patients with rightsided lesions and patients with left-sided lesions at electrode position Pz ± right-sided stimulation. (b) Grand average OERPs of healthy subjects, patients with right-sided lesions and patients with left-sided lesions at electrode position Pz ± left-sided stimulation.
1528
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
Table 2 Distribution of maximum amplitudes and amplitude values of single patients a Patients with right-sided lesions Stimulated nostril
Patients with left-sided lesions
N1P2-amplitudes Right
N1P3-amplitudes Left
Right
(A) Distribution of maximum amplitude of single patients (P 1±P 10) P2 fz, f3, f4 f3, f4 P4 f4 P8 fz, f4 fz, f3, f4 P9 f4 fz, f3, f4 P 10 f4 f4
Left
f4 f4
P P P P P
1 3 5 6 7
(B) Signi®cant changes of amplitudes: comparisons of patients (P 1±P 10) with healthy controls P2 # # P1 P4 # # # # P3 P8 P5 P9 # P6 P 10 # # # # P7
N1P2-amplitudes
N1P3-amplitudes
Right
Left
Right
f4
f4 f4 fz, f3, f4 f3
Left
fz, f3, f4 f3
# # #
#
a
fz, amplitude values at Fz . amplitude values at Pz; f3, amplitude values at F3 . amplitude values at P3; f4, amplitude values at F4 . amplitude values at P4. # , decreased
4. Discussion The present investigation indicates marked de®cits in olfactory performance in patients with supratentorial brain tumors. Results of the discrimination task showed decreased numbers of correct reactions in both patients with rightsided and patients with left-sided lesions. However, in¯uence of the side of stimulation differed in both groups: whereas patients with left-sided lesions only had an attenua-
tion of correct reactions after left-sided, i.e. ipsilateral, stimulation, patients with right-sided lesions showed reduced numbers of correct reactions after right- and leftsided stimulation. This ®nding is in accordance with the results of numerous studies (e.g. Zatorre and Jones-Gotman, 1991) that reported greater olfactory impairment in patients with lesions of the right hemisphere in comparison to patients with analogous lesions of the left hemisphere. Also, present OERP-data of patients with right-sided lesions
Fig. 4. Grand average AERPs at electrode positions Fz, Cz, and Pz in patients with right-sided lesions/left-sided lesions.
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
indicate that structural lesions of the right hemisphere may alter olfactory processing steps after right-sided and leftsided stimulation. In these patients, an alteration of topographical distribution of maximum amplitudes could be observed independent of the side of stimulation. Hummel et al. (1995) found comparable results in patients with rightsided temporal lobe epilepsy. Additionally, they reported latency differences (prolonged latencies after ipsilateral stimulation) between patients with right- or left-sided epileptic foci to be more pronounced after stimulation of the right nostril, which also might be related to more severe olfactory impairment after lesions of the right hemisphere. It is remarkable that only one of the patients in the present study had recognized an impairment of olfactory functions in advance. Doty et al. (1997) reported that many patients with olfactory loss secondary to head trauma were inaccurate in judging the state of their chemosensory functioning. Also, patients with unilateral anosmia after neurosurgical operations via the fronto-temporal approach most often did not recognize this complication (Eriksen et al., 1990). These disparities between subjective impression and objective test results seem to be characteristic for disturbances of the olfactory system and also have been described in patients with chronic neurological disease, e.g. Parkinson's disease (Doty et al., 1988). Considering special electrode positions, the patients' OERP-data showed distinct differences compared to data of the control group. In contrast, an evaluation of mean amplitude values of the 7 recording sites exhibited no significant effects. This ®nding emphasizes the importance of EEG recording from numerous electrode positions and a position-related analysis of OERP-data. In patients with right-sided lesions, decreased P2- and P3amplitudes could be found at parietal electrode positions. At the frontal recording sites, amplitudes did not show differences in comparison to the control group. In contrast to healthy subjects, where amplitudes increased from frontal to parietal sites, maximum amplitude values of patients with right-sided lesions were measured at frontal recording sites. These ®ndings compare to the observations that had been made in healthy elderly subjects as well as in patients suffering from right-sided temporal lobe epilepsy. Several authors reported age-related effects of topographical distribution of P2-amplitudes: OERP-data of the elderly (mean 66 years) in response to amyl acetate showed differences from data of young adults (mean 27 years) in P2-amplitudes at recording positions Cz and Pz, but not at the frontal site Fz (Murphy et al.,1994). In young subjects, amplitude increased from Fz to Cz to Pz; elderly subjects showed no such increase. Similar observations have been made for the P3-component (Murphy et al., 1998; Geisler et al., 1999). Hummel et al. (1998) investigated subjects of 3 age ranges (15±34, 35±54, and 55±74 years) and reported that the fronto-parietal topographical differences between the OERP-amplitudes in response to vanillin and hydrogen sul®de leveled off with increasing age. A comparable observation was made in
1529
patients with right-sided temporal lobe epilepsy (Hummel et al., 1995). In contrast to patients with a left-sided focus, who showed normal distribution of maximum P2-amplitudes, these patients elicited largest P2-amplitudes at Cz. This shift of topographical distribution of amplitudes might indicate an alteration of direction or location of the dipoles producing the OERP. A change of dipole characteristics could be expected, when brain regions that are normally involved in olfactory processing steps are destroyed by brain tumors as well as by epileptic seizures or neurodegenerative disease. The question remains whether the observed electrophysiological ®ndings re¯ect an impairment of olfactory processing or just represent an epiphenomenon of modality-non-speci®c changes of the CNS, which could be expected as a result of the brain tumor, the edema, and the destruction of neuronal structures. Our ®ndings concerning patients' impairment in the olfactory discrimination task on one hand and alteration of the P2-amplitude on the other hand would support a modality-speci®c characteristic of this OERP-component as suggested by several authors (Murphy et al., 1994; Evans et al., 1995; Hummel et al., 1998). In contrast, the observed correlation between the P3-amplitudes of OERP and AERP in patients with right-sided brain lesions after ipsilateral stimulation indicates that this component might re¯ect modality-non-speci®c processing steps in the right hemisphere. The hypothesis of a modalitynon-speci®c characteristic of the P3 was already made by Pause et al. (1996) and had been supported by following studies (Murphy et al., 1998; Geisler et al., 1999). Nevertheless, as the processing steps underlying the P3-component of the OERP are associated with olfactory stimuli, they might give indirect information about an early sensory perception of an odor, which is later processed within associative cortex areas. Based on this assumption, the alteration of topographical distribution of the P3-amplitudes in patients with right-sided brain lesions, which was observed after ipsi- and contralateral stimulation, also would re¯ect the bilateral olfactory impairment of these patients. In conclusion, investigation of OERPs in patients with brain lesions can give information on the functionality of the olfactory system. The diagnostic value of the OERPs lies in the possibility of detecting unnoticed functional olfactory de®cits, screening neurological or psychiatric patients. Nevertheless, data of single patients have to be interpreted carefully and deviant ®ndings should be seen as an indication for further clinical olfactory tests. Acknowledgements We thank K. Krauel, T. Ehmke, C. MuÈller and K. Rogalski for their technical assistance. Preliminary results from this study were presented at the symposium `Chemosensory Bioresponses in Man II', held at Erlangen, Germany, November/December 1999.
1530
C. Daniels et al. / Clinical Neurophysiology 112 (2001) 1523±1530
References Cui L, Evans WJ. Olfactory event-related potentials to amyl acetate in congenital anosmia. Electroenceph clin Neurophysiol 1997;102:303± 306. Doty RL, Brugger WE, Jurs PC, Ornolorff MA, Snyder PJ, Lowry LD. Intranasal trigeminal stimulation from odor volatiles. Psychometric responses from anosmics and normal human. Physiol Behav 1978;20:175±185. Doty RL, Deems DA, Stellar S. Olfactory dysfunction in Parkinson's desease. A general de®cit unrelated to neurologic signs, disease stage or disease duration. Neurology 1988;38:1237±1244. Doty RL, Yousem DM, Pham LT, Kreshak AA, Geckle R, Lee WW. Olfactory dysfunctions in patients with head trauma. Arch Neurol 1997;54:1131±1140. Elbert T, Lutzenberger W, Rockstroh B, Birnbaumer N. Removal of ocular artifacts from the EEG ± a biophysical approach. Electroenceph clin Neurophysiol 1985;60:455±463. Eriksen KD, Boge-Rasmussen T, Kruse-Larsen C. Anosmia following operation for cerebral aneurysms in the anterior circulation. J Neurosurg 1990;72:864±865. Eslinger PJ, Damasio AR, Van Hoesen GW. Olfactory dysfunction in man: anatomical and behavioral aspects. Brain Cogn 1982;1:259±285. Evans J, Cui L, Starr A. Olfactory event-related potentials in normal human subjects. Effects of age and gender. Electroenceph clin Neurophysiol 1995;95:293±301. Geisler MW, Morgan CD, Covington JW. Neuropsychological performance and cognitive event-related brain potentials in young and elderly adults. Clin Exp Neuropsychol 1999;211:108±126. Hummel T, Pauli E, SchuÈler P, Kettenmann B, Stefan H, Kobal G. Chemosensory event-related potentials in patients with temporal lobe epilepsy. Epilepsia 1995;361:79±85. Hummel T, Barz S, Pauli E, Kobal G. Chemosensory event-related potentials change with age. Electroenceph clin Neurophysiol 1998;108:208± 217. Kettenmann B, Hummel C, Stefan H, Kobal G. Multiple olfactory activity in the human neocortex identi®ed by magnetic source imaging. Chem Senses 1997;22:493±502. Kobal G. Elektrophysiologische Untersuchungen des menschlichen Geruchssinnes. 1st ed. Stuttgart: Thieme, 1981. Kobal G, Hummel T. Olfactory evoked potentials in humans. In: Getchell TV, Doty RL, Bartoshuk LM, Snow JB, editors. Smell and taste in health and disease, New York, NY: Raven Press, 1991. pp. 255±275. Kobal G, Hummel T. Olfactory and intranasal trigeminal event-related potentials in anosmic patients. Laryngoscope 1998;108:1033±1035. Koizuka I, Yano H, Nagahara M, Mochizuki R, Seo R, Shimada K, Kubo T,
Nogawa T. Functional imaging of the human olfactory cortex by magnetic resonance imaging. Otorhinolaryngology 1994;56:273±275. Krauel K, Pause BM, Sojka B, Schott P, Ferstl R. Attentional modulation of central odor processing. Chem Senses 1998;23:423±432. Leplow B, Pause BM, Behrens C, Sojka B, Ferstl R, Mehdorn HM. Assessment of olfactory and trigeminal perception of odors in patients with lesions of the olfactory system. Poster at the 21st meeting of the German Society of Psychophysiology and its Applications, WuÈrzburg, June 9± 12, 1993. Levy LM, Henkin RI, Lin CS, Martins D, Schellinger D. Functional MRI of human olfaction. J Comput Assist Tomogr 1997;21(6):849±856. Murphy C, Nordin S, de Wijk RA, Cain WS, Polich J. Olfactory-evoked potentials: assessment of young and elderly, and comparison to psychophysical threshold. Chem Senses 1994;19:47±56. Murphy C, Wetter S, Morgan CD, Ellison DW, Geisler MW. Age effects on central nervous system activity re¯ected in the olfactory event-related potential. Ann N Y Acad Sci 1998;855:598±607. Pause BM, Krauel K. Chemosensory event-related potentials (CSERP) as a key to the psychology of odors. Int J Psychophysiol 2000;36:105±122. Pause BM, Sojka B, Krauel K, Ferstl R. The nature of the late positive complex within the olfactory event-related potential OERP. Psychophysiology 1996;33:376±384. Pause BM, Krauel K, Sojka B, Ferstl R. Is odor processing related to oral breathing? Int J Psychophysiol 1999;32:251±260. Sobel N, Prabhakaran V, Desmond JE, Glover GH, Sullivan EV, Gabrieli JDE. A method for functional magnetic resonance imaging of olfaction. J Neurosci Methods 1997;78:115±123. Solomon GS. Anosmia in Alzheimer disease. Percept Motor Skills 1994;79:1249±1250. Wenning GK, Shephard B, Hawkes C, Petruckevitch A, Lees A, Quinn N. Olfactory function in atypical parkinsonian syndromes. Acta Neurol Scand 1995;91:247±250. Yousem DM, Geckle RJ, Bilker WB, McKeown DA, Doty RL. Posttraumatic olfactory dysfunction. MR and clinical evaluation. Am J Neuroradiol 1996;17:1171±1179. Yousem DM, Williams SCR, Howard RO, Andrew C, Simmons A, Allin M, Geckle R, Suskind D, Bullmore ET, Brammer MJ, Doty RL. Functional MR imaging during odor stimulation. Preliminary data. Radiology 1997;204:833±838. Zald DH, Pardo JV. Functional neuroimaging of the olfactory system in humans. Int J Psychophysiol 2000;36:165±181. Zatorre RJ, Jones-Gotman M. Human olfactory discrimination after unilateral frontal or temporal lobectomy. Brain 1991;114:71±84. Zatorre RJ, Jones-Gotman M, Evans AC, Meyer E. Functional localization and lateralization of human olfactory cortex. Nature 1992;360:339± 340.