Abstract

Electroencephalographical (EEG) recording studies have shown that odorants produce olfactory evoked potentials (OEPs) on the scalp surface. However, EEGs can only provide limited information about the intracerebral sources from where the OEPs are generated. By contrast, intracerebral EEG recordings enable direct examination of the electrophysiological activity from a given cerebral area. In the present study, neural activity was recorded from the amygdala of seven epileptic patients undergoing intracerebral EEG recordings prior to surgical treatment for relief of intractable seizures. Two olfactory tests were used: a passive-stimulation test consisting of the successive presentation of 12 common odorants and a suprathreshold detection test including both odorant and non-odorant stimulations. Recordings from the amygdala revealed that all odorant stimulations induced large and reproducible OEPs, whereas the non-odorant stimulations did not. It was also found that repetition of the same odorant stimulation led to a decrease in the latency of the first OEP component. This modulation, which corresponds to a faster olfactory processing, strongly suggests that the amygdala is involved in early olfactory attentional processes. In conclusion, it appears that the human amygdala discriminates the incoming information from the nasal airflow as being odorant or not and, additionally, that its speed of processing is sensitive to recent experience with an odor.

Introduction

Our knowledge of the anatomy and physiology of the olfactory system is largely based on rodent and non-human primate studies. These investigations indicate that the olfactory pathways originating in the olfactory bulb are divided into two routes: (i) the lateral olfactory tract which includes fibers projecting to the piriform, periamygdaloid and entorhinal cortices in the temporal lobe, as well as fibers ending in the olfactory tubercle and the piriform cortex in the frontal lobe (Truex and Carpenter, 1969; Nieuwenhuys et al., 1978; Carmichael et al., 1994); and (ii) the medial olfactory tract that mostly supports centrifugal efferents to the olfactory bulb. Secondary olfactory projections reach the orbitofrontal cortex directly from the piriform cortex and indirectly through connections from the olfactory tubercle and the piriform cortex to the mediodorsal nucleus of the thalamus (Krettek and Price, 1977; Carmichael and Price, 1995). In the temporal lobe, secondary olfactory projections heavily invade both the amygdala and the entorhinal cortex (Krettek and Price, 1977).

In human beings, our knowledge of the olfactory system is more limited. Several investigators (Allison and Goff, 1967; Kobal and Hummel, 1988, 1991a, b; Murphy et al., 1994; Pause et al., 1999) have recorded scalp electroencephalographical (EEG) activity in order to analyse olfactory evoked potentials (OEPs) which offer some insight into the cognitive processing of odors. The successive peaks forming OEPs, which are labeled according to their polarity and latency, have been related to various stages of cognitive processing (Lorig, 1989; 1991a,b; Pause et al., 1999). The OEP waveform is generally composed of an N1, which is a negative potential approximately occurring at a latency of 300–400 ms after the stimulus onset, followed by a P2 which is a positive potential approximately occurring at a latency of 500–600 ms. The scalp OEPs are recorded from several scalp sites and present a maximal amplitude at parietal sites. A possible approach to localize more precisely the potential generators is topographic analysis of the registered activity. However, surface potential distributions have to be interpreted cautiously because potential waveforms are not produced by a single intracranial generator. Over the last few years, magneto-encephalography has been applied to obtain a better insight into cortical regions activated after olfactory stimulation (Kettenmann et al., 1996; Sakuma et al., 1997; Kobal and Kettenmann, 2000).

In epileptic patients entering into a preoperative assessment program for possible surgical treatment, the use of stereotactic EEG (SEEG) with depth electrodes was found to contribute usefully to the purpose of localizing a suspected site of seizure onset (Talairach and Bancaud, 1973; Binnie et al., 1994). Taking advantage of this procedure, few investigators have examined SEEG activity in response to olfactory stimulations. Most of these investigations were performed during acute depth EEG procedures in the operating room (Narabayashi et al., 1963; Hughes et al., 1972; Hughes and Andy, 1979), while few others recorded SEEG activity in chronically implanted patients (Halgren et al., 1977a,b). It was found that odorants induced increased frequency of EEG rhythms in the amygdala, the hippocampus, the olfactory bulb and the olfactory tract (Narabayashi et al., 1963; Hughes et al., 1972; Halgren et al., 1977a,b; Hughes and Andy, 1979). These SEEG changes, described as spindles, were yet not systematically regarded as olfactory responses. For Halgren et al., they were thought rather to reflect breathing or vascular responses related to the blood gas content. Narabayashi et al. (Narabayashi et al., 1963) used ether inhalations in the deeply anesthetized patients subjected to stereotactic amygdalectomy in an effort to confirm the proper placement of the electrode within the amygdala. The evoked spindles were larger and sharper when recorded in the medial nuclear groups, and they had a tendency to disappear after repeated inhalations, suggesting an adaptation process. For other investigators, spindles specifically appeared for a certain class of odorants that presented similar psychophysiological and stereochemical features (Hughes et al., 1972; Hughes and Andy, 1979). This suggests that the frequency component is likely to depict an important factor in neurophysiological coding of olfactory information, and thus in discrimination within the olfactory bulb, the olfactory tractus and the amygdala. The overall data reported here reveal great divergences in interpreting the olfactory evoked spindles. In addition, studies performed during acute intracerebral EEG procedures in the operating room do not offer favorable experimental conditions because of the high stress level in subjects and the limited time available for SEEG recordings.

The aim of this study was to assess the responsiveness of the human amygdala to odorants by recording OEPs in this structure during chronic SEEG recordings in epileptic patients. These patients underwent a passive-stimulation test consisting in the presentation of 12 different odorant stimulations. No explicit cognitive task was required during this test. To assert that the recorded potentials were specifically induced by odorant stimulations and not by some mechanisms related to breathing, patients were also submitted to a suprathreshold detection test including both an odor condition and a non-odor control condition.

Material and Methods

Patients

Recordings were obtained from 14 patients (five males, nine females, 18–55 years of age) suffering from partial refractory epilepsy and implanted with a chronic SEEG electrode in the right or left amygdala. Additionally, 6–15 other electrodes were implanted in various intracerebral sites, most of which were located ipsilaterally to the suspected epileptogenic zone. The choice of the anatomical targets was based on clinical, MRI and video-EEG recording data, and was independent of the present study. None of the patients exhibited nasal pathologies, complained of any disturbances in their sense of smell or had an olfactory component to the seizures such as olfactory auras. Before the experiments began, we ensured that patients had had no epileptic seizures during the preceding 24 hours. Four patients showing epileptic discharges during the test and/or frequent interictal spikes in the amygdala (Fig. 1) and two patients showing an unclear respiratory signal precluding a reliable averaging were excluded from the study. We also eliminated one patient in whom most odorant (90%) and half of the non-odorant stimulations were followed by a long spike-wave complex similar to spontaneous interictal spike wave complexes (Fig. 2a). Therefore, data from seven patients were ultimately retained for the analysis (Table 1). All patients gave their fully informed consent for participating in the study, which did not add any invasive procedure to the depth EEG recordings performed routinely in our department.

Electrophysiological Methods

The electrode implantation procedure has been described elsewhere (Talairach and Bancaud, 1973). First, a cerebral angiography was performed in stereotactic conditions using an X-ray source located 4.85 m away from the patient's head, thus eliminating the linear enlargement due to X-ray divergence. As a consequence, the films could be used for measurements without any correction. Second, the pertinent targets were identified using the three-dimensional MRI of each patient brain, previously enlarged at scale one on slices reconstructed in the three planes of Talairach's stereotactic atlas (Talairach and Tournoux, 1988). Since MRI and angiographic images were at scale one, they could easily be superimposed, thus minimizing the risk of any damage to cerebral veins or arteries during implantation. Electrodes were implanted with the patient's head fixed in the stereotactic frame using a planar grid parallel to the midline vertical plane of Talairach's atlas. Thus, all electrode tracks were perpendicular to the midline vertical plane. Each electrode had from 10 to 15 contacts of 2 mm length, separated by 1.5 mm (Fig. 3a). The anatomical locating allowed us to determine which contacts of an electrode were inside a given structure. It consisted of the superimposition of the frontal angiographic image performed, whereas electrodes were placed to the corresponding MRI slice. The electrodes were left in place for up to 15 days if necessary. The penetration depth of each electrode was measured on frontal MRI slices from the electrode tip to the midline, which was also angiographically visualized by the sagittal sinus. Stereotactic coordinates of electrode contact positions were verified by performing a scale one skull radiography with electrodes in place superimposed to scale one angiography. In this way, each contact on MRI was localized by calculating the distance between contacts and the midline vertical plane, the anterior commissure–posterior commissure (AC–PC) horizontal plane and the vertical frontal plane traversing the posterior margin of the anterior commissure (AVC). Three coordinates were measured for each contact: x for the lateral medial axis (x = 0 for the coordinate of the midline vertical plane); y for the antero-posterior axis (y = 0 for the coordinate of ACV frontal plane) and z for the vertical axis (z = 0 for the coordinate of the AC–PC horizontal plane). In order to pool data across subjects, these coordinates were used to plot each contact in the atlas (Talairach and Tournoux, 1988).

Stimuli

Thirteen odorants selected from a previously established data bank (Royet et al., 1999) were used in the present study. Eight of them (lily, pizza, caramel, bitter almond, lavender, mustard, apricot, mint-gum) were compounds furnished by perfumery and aroma societies (Givaudan-Roure, International Flavors and Fragrances) and five others (butyric acid, isoamyl phenylacetate, guaiacol, isovaleric acid, butanol) were pure chemical compounds provided by chemical manufacturers (Aldrich, Sigma). The odorants were contained in 20 ml yellow glass bottles with screw lids in polyethylene (Fisher, Erlancourt, France). An odorant solution (10%) was obtained by diluting 0.50 ml of product in 4.50 ml of an odorless solvent (mineral oil), placed into the bottle and absorbed by compressed filaments of polypropylene. Because mustard and isovaleric acid released a very intense odor and could provoke strong trigeminal reactions, they were diluted 100 times. Odorant preparations, which were kept in a refrigerator when not used, were left to reach room temperature before the experiment began. They were changed every 3 months and were never used more than five times. For the suprathreshold detection test, butanol was chosen because of its water solubility, low toxicity, ready availability in high purity and its successful application in previous experiments (Cain et al., 1983; Cain, 1989). In addition, butanol presents neutral odor quality—familiarity and near-neutral hedonicity (Royet et al., 1999)—thus inducing no particular cognitive processing. Finally, it is a potent stimulus for the olfactory system at concentrations that have no impact on the trigeminal nerve (Murphy et al., 1990).

Experimental Procedure

The whole experiment included two distinct olfactory tests: the passive-stimulation test and the suprathreshold detection test. The passive-stimulation test included the presentation of 12 odorant (OD) stimulations (lily, pizza, caramel, bitter almond, butyric acid, isoamyl phenylacetate, lavender, guaiacol, mustard, apricot, chewing-gum and isovaleric acid). Patients were instructed to keep quiet and to focus their attention on the 12 odors they had to smell, without any particular instruction. The suprathreshold detection test included 16 trials, and patients had to decide whether the stimulations were OD or non-odorant (NOD). The eight OD stimuli were made from a solution of butanol, and the eight NOD stimuli were made from odorless solvent. For both tests, odorants were presented in a fixed pseudo-random order and with an inter-stimulus interval of 1 min that is known to be sufficient to preclude sensory adaptation in the subjects (Lawless et al., 1991). Stimulations were monorhinally administered in the nostril ipsilateral to the implanted amygdala (that is, to the suspected epileptogenic side), while patients held the contralateral nostril closed. The presentation time varied from 3 to 5 s, depending upon the patients' inspiratory phase duration. During the recordings, the patients lay relaxed on their beds and their rooms were ventilated. Patients were preferentially submitted to the first olfactory test on the fifth day and to the second test on the seventh day following electrode implantations in order to minimize the acute SEEG effects of the surgery and the post-ictal SEEG changes normally observed in the second week. Before testing, patients were instructed in how to smell the odorants in order to minimize the intra-and inter-subject breathing pattern variability. Patients were trained to attend the following instructions: (i) breathe out deeply through the mouth; (ii) block up the non-stimulated nostril with a finger (as a ‘ready’ signal); and (iii) continuously inspire once the glass bottle containing the odorant was under the stimulated nostril (for the training, the glass bottle did not contain any odorants). Thus, all patients presented a steady respiratory rhythm during the stimulations and the experimenter could synchronize stimulations with the patients' breathing.

Data Acquisition and Analysis

SEEG activity was continuously recorded from up to 96 contacts using a 128 Hz sampling rate with an analog filter band-pass of 1–64 Hz (Micromed, Italy). A low-pressure sensor simultaneously monitored the phase of the respiratory cycle, while a video synchronized to SEEG recording data monitored the procedure. To average responses to the successive stimuli, time of onset of the respiratory signal was taken as time of onset of the stimulation (Fig. 2b). The breathing signal waveform contained (i) a flat line corresponding to the expiration through the mouth and to the nostril block-up time and (ii) oscillations corresponding to the inspiratory phase, i.e. to the stimulation time. The stimulus onset was then determined by the end of the flat line, that is at the beginning of the first deflection (Fig. 2b). Additionally to the breathing signals, stimulus onsets were assigned by watching video recordings that were synchronized to SEEG recordings. Analyses were assessed on monopolar recordings collected from the contact of the amygdala electrode where peak amplitudes were the largest and referenced to an intracerebral electrode contact with a ground electrode located on a different intracerebral electrode. The OEP waveform was obtained by averaging off-line the various olfactory trials for each test and for each patient. Two reproducible components were identified consisting in a first positive peak (PP) and a second negative peak (NP) followed by a slow reversion to baseline terminating at the end-point. To observe eventual polarity reversal, bipolar recordings were additionally obtained off-line by digital subtraction. Peak and end-point latencies and peak amplitudes with respect to the baseline were calculated on the raw data (non-averaged recordings) for each test and for each patient. For the eight NOD trials of the suprathreshold detection test, maximal (MAX) and minimal (MIN) amplitudes of the SEEG signal were determined from raw recordings over time-windows determined by the raw peak latencies (200–400 ms and 500–700 ms, respectively). All these data were submitted to separate one-way analysis of variance (ANOVA) with repeated measurements (Winer, 1962).

Results

OEPs Recorded from the Amygdala

Passive Stimulation Test

The 12 olfactory stimulations of the test consistently induced OEPs in the amygdala (Figs 3 and 4). These electrophysiological responses were only collected from the contacts implanted inside the amygdala (generally the four deepest contacts) and not from the other contacts of the electrode. Also, these responses were never observed in other recorded structures, including the very close ones such as the hippocampus or temporal pole (see Fig. 1). The amygdalian OEPs were large and observed in raw recordings, that is in response to a single odorant (Fig. 3b), as well as in averaged data (Figs 3c and 4a). The intracerebral potential waveform, which consisted of a PP, an NP and a subsequent slow reversion to the baseline terminating at the endpoint, was similar to that previously described for scalp potentials (Allison and Goff, 1967; Kobal and Hummel, 1988, 1991a, b; Murphy et al., 1994; Pause et al., 1999). The components exhibited similar latencies on all contacts of the amygdala electrode. In five patients, the monopolar recordings revealed that peak amplitudes progressively decreased from the deepest to the most external contact of the electrode (Fig. 3b,c). In the two other patients, the largest peak amplitude was observed on the second deepest contact, and a polarity reversal of the PP/NP response was observed in bipolar recordings between the second and the third deepest contacts of the electrode (Fig. 4), suggesting that these contacts were the closest to the source of OEPs. However, this finding does not enable us to give a clear indication of a dipolar source with regards to that of the electrode track. Mean latencies and amplitudes of the different OEP components are given in Table 2. ANOVAs did not show a significant effect of the odor factor on the PP [F(11,66) = 0.4, n.s.], NP [F(11,66) = 0.76, n.s.] and endpoint [F(11,66) = 1.55, n.s.] latencies and on the PP [F(11,66) = 1.34, n.s.] and NP [F(11,66) = 1.52, n.s.] amplitudes.

Suprathreshold Detection Test

In the suprathreshold detection test, behavioral performances were excellent since the accuracy rate reached 98.2%. Raw and averaged recordings from the amygdala showed that the OD stimulations of butanol induced high amplitude and very reproducible OEPs (Fig. 5). As previously observed, none of the other recorded structures produced similar electrophysiological activity in response to olfactory stimulations. The OEP waveform, consisting of an PP/NP pattern, was similar to that described in the preceding olfactory test. Mean amplitudes collected for the PP and NP components are given in Table 2. ANOVAs showed that PP and NP amplitudes collected after OD stimulations were significantly larger than the MAX and MIN amplitudes collected after NOD stimulations at similar latencies [F(1,12) =7.35, P < 0.025; F(1,12) =18.6, P < 0.001, respectively]. Thus, only OD stimulations, but not NOD stimulations, induced high amplitude and reproducible OEPs. ANOVAs did not reveal significant variations of the NP and PP amplitudes as a function of the OD stimulation order of presentation (from 1 to 8) [F(7,42) = 1.52, n.s.; F(7,42) = 2.28, n.s., respectively]. Mean latencies collected for the PP, NP and endpoint components are given in Table 2. Mean PP latencies of OEPs were calculated as a function of eight OD stimulations and are depicted in Figure 6. ANOVA revealed a significant effect of repeated butanol stimulation on the PP latency [F(7,42) = 2.68, P < 0.025], but not on the NP and endpoint latencies [F(7,42) = 2.06, n.s.; F(7,42) = 1.06, n.s., respectively]. Multiple orthogonal comparisons between pairs of means showed that the PP latency of OEPs induced by the first OD stimulation was significantly longer (P < 0.05) than those induced by five other OD stimulations (the 5th, 7th, 11th, 12th and 15th stimulations). They also show that the PP latency of OEPs induced by the 7th OD stimulation was significantly smaller (P < 0.05) than that produced by the 10th OD stimulation (Fig. 6).

Evoked Spindles Recorded from the Amygdala

While OEPs were exclusively recorded in response to odorant stimulations, spindles were observed in both OD and NOD conditions (Fig. 7). Spindle latencies and durations were extremely variable between odorants and between patients. Spindles generally appeared after the OEPs in the OD conditions or at a similar latency in the NOD condition with respect to the time onset of stimulation. They generally lasted until the end of the stimulation, that is until the end of the inspiratory phase.

Discussion

In the present study, the neural activity of various intracerebral structures implanted with an electrode was recorded in epileptic patients submitted to olfactory testing. Our data show that the human amygdala is widely and reliably responsive to odorant stimulations, since this latter structure was found to produce reliable and large OEPs. These OEPs were characterized by a positive potential with a latency of 300 ms after the stimulus onset, and by a negative potential with a latency of 450 ms after the stimulus onset. These components exhibited similar latencies to those recorded from the amygdala in other sensory modalities (Seeck et al., 1993, 1995; Guillem et al., 1998) and have never been described in previous olfactory depth recording studies.

Recordings from the Amygdala

Firstly, our results demonstrate that OD stimulations induced OEPs in the amygdala, whereas NOD stimulations were ineffective in eliciting such electrophysiological responses. The possibility that OEPs were due to breathing mechanisms is thus unlikely. This result is supported by the fact that in the suprathreshold detection test the motor and sensory aspects, regardless of the olfactory ones, were presumably identical for OD and NOD stimulations. Thus, the OEPs recorded from the amygdala appear to be specific responses to odorant stimulations. Secondly, our results showed that both OD and NOD stimulations elicited spindles in the amygdala. This non-specific increase of the firing rate has previously been reported as resulting from cardiovascular and respiratory regulations (Halgren et al., 1977a,b). From our findings, it thus appears that two types of electrophysiological events can be distinguished in the amygdala response to an olfactory stimulation: (i) the OEPs that are specific odor-induced activations and (ii) the spindles that are probably linked to breathing (Halgren et al., 1977b). This dissociation suggests that the human amygdala discriminates the incoming information from the nasal airflow as being odorant or not.

Because most odorants stimulate both the olfactory and trigeminal nerves in the nasal cavity (Cain, 1974; Silver and Moulton, 1982; Silver et al., 1985; Bouvet et al., 1987), scalp OEPs are assumed to stem from both olfactory and trigeminal activations (Allison and Goff, 1967; Kobal and Hummel, 1988, 1991b; Hummel and Kobal, 1999; Geisler and Murphy, 2000). Although some studies demonstrated that the odorant-induced activation is primarily evoked by stimulation of olfactory receptors rather than nasal trigeminal afferents (Allison and Goff, 1967; Geister and Murphy, 2000), it remains difficult to distinguish the respective contribution of each system. Nevertheless, we must note two major observations made within the framework of the present study. First, if trigeminal stimuli induced somatosensory evoked potentials in the amygdala, we could suppose their amplitudes to be higher than those of OEPs, given the intrinsic strength of trigeminal stimuli. No such result was observed in the current study, despite the fact we used such stimuli as isovaleric acid and mustard, known to induce trigeminal reactions. Second, we must also emphasize that olfactory projections onto the amygdala are bisynaptic and dense (Shipley and Reyes, 1991; Carmichael et al., 1994; Swanson and Petrovich, 1998). Swanson and Petrovich (Swanson and Petrovich, 1998) distinguished 13 structural areas in the rat amygdala, of which ten constituted the corticomedial group and were more specifically involved in olfactory functions. In man, Crosby et al. (Crosby et al., 1962) reported that five out of eight amygdaloid areas pertain to the corticomedial group. In summary, although we cannot draw irrefutable conclusions in the present study about the respective parts played by each sensory system in the generation of evoked potentials in the amygdala, the bulk of evidence suggests that these potentials are likely to derive from the olfactory pathways.

As suggested above, the human amygdala can be separated into two major division: a dorsomedial group and a basolateral group. The dorsomedial group is constituted of nuclei localized in the dorsal, but also in the medial part of the amygdala. The basolateral nuclear complex represents the largest and the better-differentiated part of the human amygdaloid complex. Due to the horizontal orientation of electrodes, we can suppose that contacts were mainly situated in the basolateral group. However, the deepest contacts were even closer to it and could also be in areas belonging to the dorsomedial group, such as the cortical amygdaloid nucleus and the cortico-amygdaloid transition area, since the less deep contacts were probably localized in the lateral, the basal (pars lateralis), the accessory basal, the central amygdaloid nuclei. As a consequence, maximal amplitudes of OEPs found with the deepest contacts can be explained by activities more directly related to processing of olfactory information.

That none of the previous SEEG studies reported OEP in the human amygdala could be due to experimental conditions such as electrode impedance, electrode positions, stimulation conditions, etc. Thus, for instance, Halgren et al. (Halgren et al., 1977a,b) did not observe response to passive olfactory and trigeminal stimulations, but used microelectrodes (7–9-strand bundles of fine wires of 40 μm diameter) inserted with cannulae. In the present paper, we used electrodes comprising from 10 to 15 contacts of 2 mm length. Such differences in methodology do not allow for picking up the same electrophysiological phenomena.

Such large OEPs have also never been reported in magneto-encephalographic studies, probably for of two main reasons. Firstly, the human amygdala is quite a small structure buried inside the medial part of the temporal lobes, under the brain and, indeed rather inaccessible to this method. Second, the amygdala is a nuclear structure that corresponds, in accordance with a previous model of potential field generators (Lorente de Nó, 1947), to a closed-field structure in which the cell somas are gathered in the center of the structure and the dendrites extend radially outwards. Somatic depolarization produces an inward current and maximal negative potential at the center of the nucleus. This radial inward current causes the potential gradient to be zero anywhere outside the nucleus. Consequently, the electrical activity cannot be recorded using electrodes situated at the periphery of the nucleus. This latter model agrees with our data, since large OEPs were observed only from the discrete recording sites of the electrode located in the amygdala (A1–A4 in Figs 3 and 4) and not from those located outside in the white matter or in the cortex. On the whole, these considerations could explain the impossibility of observing such huge responses in the amygdala with magneto-encephalography.

OEPs Variability

Our results do not show evidence of an effect of psychophysiological dimensions of the odorants, such as intensity or trigeminal action, which are yet known to modulate the scalp EEG potential (Kobal and Hummel, 1988, 1991a,b). This may be attributed to the important variability of the OEP latencies and amplitudes revealed by the raw recordings as a function of odorants and patients. In a previous scalp EEG recording study, this variability has been attributed to factors related to subjects such as olfactory sensitivity, sex or age (Murphy et al., 1994; Evans et al., 1995; Hummel et al., 1998). In the present study, it could possibly be emphasized by additional factors such as: (i) the various anti-epileptic drugs taken by patients that could differentially modulate global cerebral activity and affect neuron responsiveness during olfactory testing; (ii) the functional lesions presented by the epileptic patients that may differentially alter the processing of olfactory information (Hummel et al., 1995; Savic et al., 1997); (iii) the inter-hemispheric differences of the amygdala electrode position; and (iv) the various positions of electrodes inside the amygdala (Narabayashi et al., 1963). All these factors involved in the modulation of the peak latencies and amplitudes were not controlled for and may have interfered with the factors linked to the psychophysiological dimensions of the odorants.

Amygdala Sensitivity to Selective Attention Effects

Amygdala neural activity was found to be strongly and dynamically influenced by recent experience with an odor, producing OEPs with various PP component latencies. This observation was only obtained in the suprathreshold detection test for which all OD stimulations were identical. The PP latency obtained to the first stimulation of butanol was longer than those obtained for the seven subsequent stimulations and tended to decrease progressively, except when two NOD stimulations were inserted. Such electrophysiological findings have not been reported in previous intracerebral recordings from the amygdala. In scalp EEG recording studies, the effect of stimulus repetition upon chemosensory evoked potentials was investigated in order to assess the habituation phenomenon (1991a,b; Hummel and Kobal, 1999). While peak amplitudes decreased with repeated stimulations, peak latencies remained unchanged. Thus, based on these results, we can not assume that habituation contributed to the latency reductions observed in the present study. By contrast, we could hypothesize that they might be due to a selective attention phenomenon, that is to stimulus predictability. Selective attention is an essential cognitive function that enables us preferentially to select and process high priority information, thus improving perception by modifying sensory inputs at an early stage of processing (Egeth, 1967). In auditory and visual modalities, selective attention was assessed by recording scalp potentials in human subjects (Näätänen and Michie, 1979; Hillyard and Mangun, 1987). The attention effect was found to modulate the early peak amplitudes, expressing a sensory gating of the neural generators which increased their level of activation for attended stimuli and decreased it for not attended stimuli. Thus, this is the amount of the sensory gain that is reported to be modified. More recently, research on the selective attention effect on the first component of the evoked potentials (N1) was reviewed in the olfactory modality (Krauel et al., 1998; Pause et al., 1999). By contrast to the auditory and visual modalities, the sensory gating expressed its effect on the N100 latencies rather than on their amplitudes. These authors concluded that the speed of olfactory processing was enhanced by selective attention. From this, we suggest that this process of time-dependent facilitation of neural transmission may also explain our results.

Role of the Amygdala in Odor Processing

Participation of the human amygdala in processing olfactory information is consistent with data showing olfactory deficits in epileptic patients with unilateral cerebral excision from temporal lobes, or non-operated patients suffering from temporal epilepsy. Although discrepancies were noted, especially for odor detection and recognition threshold (Huber et al., 1965; Rausch and Serafetinides, 1975a,b; Henkin et al., 1977; Eichenbaum et al., 1983; Jones-Gotman and Zatorre, 1988), olfactory performances were mainly found to be impaired in studies with respect to quality discrimination, recognition memory, matching task, and identification (Rausch et al., 1977; Abraham and Mathaï, 1983; Eichenbaum et al., 1983; Eskenazi et al., 1983, 1986; Zatorre and Jones-Gotman, 1991; Carroll et al., 1993; Martinez et al., 1993; Hummel et al., 1995; Savic et al., 1997; Jones-Gotman et al., 1997). Except that amygdalectomy abolishes olfactory areas (Chitanondh, 1966; Andy, 1967; Andy et al., 1975), none of the studies including patients with a lesion specifically circumscribed to the amygdala reported olfactory performance impairments. The observation of OEPs induced in the amygdala by the odors used in the present study thus allows us to suppose that patients performed a level of olfactory processing probably higher than a simple odor detection task. In such a case, OEPs could partly reflect a top–down process. These data are consistent with psychophysical studies (Stuiver, 1958) showing human odor detection reaction times shorter (180 ms) than OEP latencies obtained in the present study (300–400 ms for PP and 500–600 ms for NP). These latencies approximate more closely the reaction times observed in an odor recognition task (Laing and MacLeod, 1992). This suggestion is also reinforced by the fact that event-related potentials can also be recorded in the amygdala in other sensory modalities (Seeck et al., 1993, 1995; Guillem et al., 1998). Indeed, incoming visual or auditory information does not directly converge into the amygdala, as for olfactory information, but must first be conveyed through the thalamus. In recent cerebral imaging studies, the amygdala was shown as being strongly involved in the processing of emotional responses and in the hedonic judgement of odorants (Zald and Pardo, 1997; Zald et al., 1998; Royet et al., 2000a,b). We can suppose that OEPs observed in the present study could also partly reflect the emotional component of stimuli. Further studies are needed to evidence possible roles of the amygdala in olfactory processing using SEEG recordings.

Conclusion

In the study of epilepsy, the analysis of SEEG signals recorded with depth electrodes in the amygdala provides major information on olfactory processing. The present study shows that different types of electrophysiological responses occur in the amygdala after odorant stimulation, but that only the evoked potentials are olfactory responses. The finding that the human amygdala was responsive to olfactory stimulation provides a firm basis for further exploration of the neural bases of olfactory processing.

Notes

We are grateful to the staff of the Unité de Neurologie Fonctionnelle et d'Epileptologie for assistance in the electrophysiological recordings, to Dr Jean Isnard and Dr Catherine Fisher for their cooperation and for providing information on patients, and to Dr Marc Guénot for stereotactic electrode implantations. We thank Michel Vigouroux, Vincent Farget and Bernard Bertrand for the low-pressure sensor design, and we are very indebted to Dr Liliane Astic for critical discussions and to anonymous reviewers for helpful comments on a draft of this manuscript. This work was supported by research grants from the Region Rhône-Alpes, the Gouvernement d'Intérêt Scientifique (Sciences de la Cognition), the Centre National de la Recherche Scientifique (CNRS), the University Claude-Bernard—Lyon 1 and the Roudnitska Foundation.

Address correspondence to Julie Hudry, Laboratoire de Neurosciences et Systèmes Sensoriels, Université Claude-Bernard—Lyon 1, 50 Avenue Tony Garnier, F-69366 Lyon Cedex 07, France. Email: hudry@olfac.univ-lyon1.fr.

Table 1

Characteristics of the epileptic patients

Patient Gender Age (years) Handedness Hemispheric predominance for language Age of seizure onset (years) Seizure focus 
aPresence of an electrode in the right amygdala. 
bPresence of an electrode in the left amygdala. 
cThe WADA test showed a right hemispheric predominance for language comprehension and a left hemispheric predominance for language expression. 
P1a female 24 right left right temporal 
P2a female 48 right left 26 right fronto-temporal 
P3b male 32 right left 12 left temporal 
P4a female 35 left nonec right temporal 
P5a female 30 right left 22 right temporal 
P6b female 27 right left 12 left temporal 
P7a female 34 right left right temporal 
Patient Gender Age (years) Handedness Hemispheric predominance for language Age of seizure onset (years) Seizure focus 
aPresence of an electrode in the right amygdala. 
bPresence of an electrode in the left amygdala. 
cThe WADA test showed a right hemispheric predominance for language comprehension and a left hemispheric predominance for language expression. 
P1a female 24 right left right temporal 
P2a female 48 right left 26 right fronto-temporal 
P3b male 32 right left 12 left temporal 
P4a female 35 left nonec right temporal 
P5a female 30 right left 22 right temporal 
P6b female 27 right left 12 left temporal 
P7a female 34 right left right temporal 
Table 2

Latencies and amplitudes of the OEPs

 Latencies (ms) Amplitudes (μV) 
 PP NP Endpoint PP NP 
Passive-stimulation test 302.8 ± 139.0 467.1 ± 89.5 818.9 ± 227.8 290.1 ± 139.0 154.9 ± 91.9 
Suprathreshold detection test 300.5 ± 94.4 452.0 ± 137.3 758.1 ± 217.1 237.9 ± 167.6 158.0 ± 80.6 
 Latencies (ms) Amplitudes (μV) 
 PP NP Endpoint PP NP 
Passive-stimulation test 302.8 ± 139.0 467.1 ± 89.5 818.9 ± 227.8 290.1 ± 139.0 154.9 ± 91.9 
Suprathreshold detection test 300.5 ± 94.4 452.0 ± 137.3 758.1 ± 217.1 237.9 ± 167.6 158.0 ± 80.6 
Figure 1.

 Example of raw monopolar recordings collected from the most external contact of electrodes implanted in the left amygdala (Am), the left hippocampus (Hi) and the left temporal pole (Tp). The recordings were obtained from a patient with left temporal epilepsy during the suprathreshold detection test in response to a stimulation of butanol. The amygdala produced too frequent interictal spikes so that no OEP could be evidenced. The hippocampus and the temporal pole that did not produce epileptiform activity did not exhibit an electrophysiological response to the stimulation of butanol.

Figure 1.

 Example of raw monopolar recordings collected from the most external contact of electrodes implanted in the left amygdala (Am), the left hippocampus (Hi) and the left temporal pole (Tp). The recordings were obtained from a patient with left temporal epilepsy during the suprathreshold detection test in response to a stimulation of butanol. The amygdala produced too frequent interictal spikes so that no OEP could be evidenced. The hippocampus and the temporal pole that did not produce epileptiform activity did not exhibit an electrophysiological response to the stimulation of butanol.

Figure 2.

 Example of raw monopolar recording obtained in one patient suffering from left temporal epilepsy. (a) Long and large spike-wave complex induced by the odor of lily and collected on the electrode implanted in the left amygdala. The most external contact of the electrode is here represented. (b) Respiratory signal simultaneously collected by the low-pressure sensor enabling the determination of the onset and the end of the stimulation.

Figure 2.

 Example of raw monopolar recording obtained in one patient suffering from left temporal epilepsy. (a) Long and large spike-wave complex induced by the odor of lily and collected on the electrode implanted in the left amygdala. The most external contact of the electrode is here represented. (b) Respiratory signal simultaneously collected by the low-pressure sensor enabling the determination of the onset and the end of the stimulation.

Figure 3.

 Electrophysiological recordings from the amygdala of Patient 4 for the passive-stimulation test. (a) Frontal MRI showing the electrode implanted in the right amygdala. Only the four deepest contacts (A1, A2, A3, A4) of this electrode were considered as being located within the structure. (b) Example of raw monopolar OEPs (filtered 1–64 Hz) obtained from A1, A2, A3, A4 in response to a single stimulation of lavender. (c) Monopolar averaged OEPs (filtered 1–64 Hz) obtained from the 12 odors. PP, positive potential; NP, negative potential. Stimulus onset corresponds to the value 0 ms.

Figure 3.

 Electrophysiological recordings from the amygdala of Patient 4 for the passive-stimulation test. (a) Frontal MRI showing the electrode implanted in the right amygdala. Only the four deepest contacts (A1, A2, A3, A4) of this electrode were considered as being located within the structure. (b) Example of raw monopolar OEPs (filtered 1–64 Hz) obtained from A1, A2, A3, A4 in response to a single stimulation of lavender. (c) Monopolar averaged OEPs (filtered 1–64 Hz) obtained from the 12 odors. PP, positive potential; NP, negative potential. Stimulus onset corresponds to the value 0 ms.

Figure 4.

 Electrophysiological recordings from the amygdala of Patient 2 for the passive-stimulation test. (a) Monopolar averaged OEPs (filtered 1–64 Hz) obtained from the 12 odors on the four deepest contacts (A1, A2, A3, A4) of the electrode implanted in the amygdala. (b) Bipolar averaged OEPs (filtered 1–64 Hz) obtained by digital subtraction of the monopolar data. Stimulus onset corresponds to the value 0 ms.

Figure 4.

 Electrophysiological recordings from the amygdala of Patient 2 for the passive-stimulation test. (a) Monopolar averaged OEPs (filtered 1–64 Hz) obtained from the 12 odors on the four deepest contacts (A1, A2, A3, A4) of the electrode implanted in the amygdala. (b) Bipolar averaged OEPs (filtered 1–64 Hz) obtained by digital subtraction of the monopolar data. Stimulus onset corresponds to the value 0 ms.

Figure 5.

 Representative monopolar recordings obtained from the deepest electrode contact (filtered 1–64 Hz) in the amygdala of Patient 6 during the suprathreshold detection test. (a) Raw recordings collected in response to the first and second odorant stimulations of butanol (S1, S4), and to the first non-odorant stimulation (S2). (b) Averaged recordings for eight OD stimulations of butanol and eight NOD stimulations. Stimulus onset corresponds to the value 0 ms.

Figure 5.

 Representative monopolar recordings obtained from the deepest electrode contact (filtered 1–64 Hz) in the amygdala of Patient 6 during the suprathreshold detection test. (a) Raw recordings collected in response to the first and second odorant stimulations of butanol (S1, S4), and to the first non-odorant stimulation (S2). (b) Averaged recordings for eight OD stimulations of butanol and eight NOD stimulations. Stimulus onset corresponds to the value 0 ms.

Figure 6.

 Latencies of the OEP positive potential averaged for the seven patients and for the eight OD stimulations (S1, S4, S5, S7, S10, S11, S12, S15) of the suprathreshold detection test. *Significant decrease (P < 0.05) in latency when compared with the first stimulation; **significant difference (P < 0.05); vertical bars, standard deviation.

Figure 6.

 Latencies of the OEP positive potential averaged for the seven patients and for the eight OD stimulations (S1, S4, S5, S7, S10, S11, S12, S15) of the suprathreshold detection test. *Significant decrease (P < 0.05) in latency when compared with the first stimulation; **significant difference (P < 0.05); vertical bars, standard deviation.

Figure 7.

 Spindles recorded from raw electrophysiological responses (filtered 20–64 Hz) in the amygdala in (a) the passive-stimulation test for Patient 1 stimulated by isoamyl phenylacetate (ISO), and for Patient 3 stimulated by lavender (LAV); (b) the suprathreshold detection test for Patients 1 and 3 stimulated by an OD stimulus (butanol) and by an NOD stimulus. Stimulation duration was determined by the respiratory signal. Stimulus onset corresponds to the value 0 ms.

Figure 7.

 Spindles recorded from raw electrophysiological responses (filtered 20–64 Hz) in the amygdala in (a) the passive-stimulation test for Patient 1 stimulated by isoamyl phenylacetate (ISO), and for Patient 3 stimulated by lavender (LAV); (b) the suprathreshold detection test for Patients 1 and 3 stimulated by an OD stimulus (butanol) and by an NOD stimulus. Stimulation duration was determined by the respiratory signal. Stimulus onset corresponds to the value 0 ms.

References

Abraham A, Mathai KV (
1983
) The effect of right temporal lobe lesions on matching of smells.
Neuropsychologia
 
21
:
277
–281.
Allison T, Goff WR (
1967
) Human cerebral evoked potentials to odorous stimuli.
Electroencephalogr Clin Neurophysiol
 
14
:
331
–343.
Andy OJ (
1967
) The amygdala and hippocampus in olfactory aura.
EEG Clin Neurophysiol
 
23
:
291
–293.
Andy OJ, Jurko MF, Hughes JR (
1975
) The amygdala in relation to olfaction.
Confin Neurol
 
37
:
215
–222.
Binnie CD, Elwes RD, Polkey CE, Volans AAD (
1994
) Utility of stereo-electroencephalography in preoperative assessment of temporal lobe epilepsy.
J Neurol Neurosurg Psychiat
 
57
:
58
–65.
Bouvet JF, Gordinot F, Croze S, Delaleu JC (
1987
) Olfactory receptor cell function is affected by trigeminal nerve activity.
Neurosci Lett
 
77
:
181
–186.
Cain WS (
1974
) Contribution of the trigeminal nerve to perceived odor magnitude.
Ann NY Acad Sci
 
237
:
28
–34.
Cain WS (
1989
) Testing olfaction in a clinical setting.
Ear Nose Throat J
 
68
:
316
–328.
Cain WS, Gent J, Catalanotto FA, Goodspeed RB (
1983
) Clinical evaluation of olfaction.
Am J Otolaryngol
 
4
:
252
–256.
Carmichael ST, Price JL (
1995
) Limbic connections of the orbital and medial prefrontal cortex in macaque monkeys.
J Comp Neurol
 
363
:
615
–641.
Carmichael ST, Clugnet M-C, Price JL (
1994
) Central olfactory connections in the macaque monkey.
J Comp Neurol
 
346
:
403
–434.
Carroll B, Richardson JTE, Thompson P (
1993
) Olfactory information processing and temporal lobe epilepsy.
Brain Cognit
 
22
:
230
–243.
Chitanondh H (
1966
) Stereotaxic amygdalotomy in the treatment of olfactory seizures and psychiatric disorders with olfactory hallucination.
Confin Neurol
 
27
:
181
–196.
Crosby EC, Humphrey T, Lauer EW (1962) Correlative anatomy of the nervous system. Ch. 7: Telencephalon; part 2: subcortical telencephalic nuclei, pp. 356–393. New York: Macmillan
Egeth H (
1967
) Selective attention.
Psychol Bull
 
67
:
41
–57.
Eichenbaum H, Morton TH, Potter H, Corkin S (
1983
) Selective olfactory deficits in case H.M.
Brain
 
106
:
459
–472.
Eskenazi B, Cain WS, Novelly RA, Friend KB (
1983
) Olfactory functioning in temporal lobectomy patients.
Neuropsychologia
 
21
:
365
–374.
Eskenazi B, Cain WS, Friend K (
1986
) Exploration of olfactory aptitude.
Bull Psychon Soc
 
24
:
203
–206.
Evans WJ, Cui L, Starr A (
1995
) Olfactory event-related potentials in normal subjects: effects of age and gender.
Electroencephalogr Clin Neurophysiol
 
9
:
293
–301.
Geisler MW, Murphy C (
2000
) Event-related potentials to attended and ignored olfactory and trigeminal stimuli.
Int J Psychophysiol
 
37
:
309
–315.
Guillem F, N'Kaoua B, Rougier A, Claverie B (
1998
) Location of the epileptic zone and its physiopathological effects on memory-related activity of temporal lobe structures: a study with intracranial event-related potentials.
Epilepsia
 
39
:
928
–941.
Halgren E, Babb TL, Rausch R, Crandall PH (
1977
) Neurons in the human basolateral amygdala and hippocampal formation do not respond to odors.
Neurosci Lett
 
4
:
331
–335.
Halgren E, Babb TL, Crandall PH (
1977
) Responses of human limbic neurons to odorants induced changes in blood gases.
Brain Res
 
132
:
43
–63.
Henkin RI, Comiter H, Fedio P, O'Doherty D (
1977
) Deficits in taste and smell recognition following temporal lobectomy.
Trans Am Neurol Assoc
 
102
:
146
–150.
Hillyard SA, Mangun GR (
1987
) Sensory gating as a physiological mechanism for visual selective attention.
Electroencephalogr Clin Neurophysiol
 
40
Suppl:
61
–67.
Huber Z, Pruszewicz A, Szmeja Z, Bialek E (
1965
) Study of smell, taste, hearing, balance, sight, and tactile sensation following excision of the anterior temporal lobe.
Pol Neurol Neurosurg Psychiat
 
15
:
475
–480.
Hughes JR, Andy OJ (
1979
) The human amygdala. I: Electrophysiological responses to odorants.
Electroencephalogr Clin Neurophysiol
 
46
:
428
–443.
Hughes JR, Hendrix DE, Andy OJ, Wang C, Peeler D, Wetzel N (1972) Correlations between electrophysiological and subjective responses to odorants as recorded from the olfactory bulb, tract and amygdala of waking man. In: Neurophysiology studied in man (Somjen G, ed.), pp. 260–280. Amsterdam: Excerpta Medica.
Hummel T, Kobal G (
1999
) Chemosensory event-related potentials to trigeminal stimuli change in relation to the interval between repetitive stimulation of the nasal mucosa.
Eur Arch Otorhinolaryngol
 
256
:
16
–21.
Hummel T, Pauli E, Schüler P, Kettenmann B, Stefan H, Kobal G (
1995
) Chemosensory event-related potentials in patients with temporal lobe epilepsy.
Epilepsia
 
36
:
79
–85.
Hummel T, Barz S, Pauli E, Kobal K (
1998
) Chemosensory event-related potentials change with age.
Electroencephalogr Clin Neurophysiol
 
108
:
208
–217.
Jones-Gotman M, Zatorre RJ (
1988
) Olfactory identification deficits in patients with focal cerebral excision.
Neuropsychologia
 
26
:
387
–400.
Jones-Gotman M, Zatorre RJ, Cendes F, Olivier A, Andermann F, McMackin D, Staunton H, Siegel AM, Wieser HG (
1997
) Contribution of medial versus lateral temporal-lobe structures to human odour identification.
Brain
 
120
:
1845
–1856.
Kobal G, Hummel C (
1988
) Cerebral chemosensory evoked potentials elicited by chemical stimulation of the human olfactory and respiratory nasal mucosa.
Electroencephalogr Clin Neurophysiol
 
71
:
241
–250.
Kobal G, Hummel T (1991a) Olfactory evoked potentials in human. In: Smell and taste in health and disease (Getchell TV, ed.), pp. 255–275. New York: Raven Press.
Kobal G, Hummel T (1991b) Human electro-encephalograms and brain responses to olfactory stimulation. In: The human sense of smell (Laing DG, Doty RL, Breipohl W, eds), pp. 135–151. Berlin: Springer.
Kobal G, Kettenmann B (
2000
) Olfactory functional imaging and physiology.
Int J Psychophysiol
 
36
:
157
–163.
Krauel K, Pause BM, Sojka B, Schott P, Ferstl R (
1998
) Attentional modulation of central odor processing.
Chem Senses
 
23
:
423
–432.
Krettek JE, Price JL (
1977
) Projections from the amygdaloid complex to the cerebral cortex and thalamus in the rat and cat.
J Comp Neurol
 
172
:
687
–722.
Laing DG, MacLeod P (
1992
) Reaction time for the recognition of odor quality.
Chem Senses
 
17
:
337
–346.
Lawless HT, Glatter S, Hohn C (
1991
) Context-dependent changes in the perception of odour quality.
Chem Senses
 
16
:
349
–360.
Lorente de Nó R (
1947
) Analysis of the distribution of action currents of nerve in volume conductors.
Stud Rockefeller Inst Med Res
 
132
:
384
–477.
Lorig TS (
1989
) Human EEG and odor response.
Prog Neurobiol
 
33
:
387
–398.
Martinez BA, Cain WS, De Wijk PA, Spencer DD, Novelty RA, Sass KJ (
1993
) Olfactory functioning before and after temporal lobe resection for intractable seizures.
Neuropsychology
 
7
:
351
–363.
Murphy C, Gilmore MM, Seery CS, Salmon DP, Lasker BP (
1990
) Olfactory thresholds are associated with degree of dementia in Alzheimer's disease.
Neurobiol Aging
 
11
:
465
–469.
Murphy C, Nordin S, De Wijk RA, Cain WS, Polich J (
1994
) Olfactory evoked potentials: assessment of young and elderly, and comparison to psychophysiological threshold.
Chem Senses
 
19
:
47
–56.
Näätänen R, Michie PT (
1979
) Early selective-attention on the evoked potential: a critical review and reinterpretation.
Biol Psychol
 
8
:
81
–136.
Narabayashi H, Nagao T, Saito Y, Yoshida M, Nagahata M (
1963
) Stereotactic amygdalectomy for behaviour disorders.
Arch Neurol
 
9
:
1
–16.
Nieuwenhuys R, Voogd J, Van Huijzen C (1978) Olfactory and Limbic system. In: The human central nervous system: a synopsis and atlas, pp. 181–215. Berlin: Springer Verlag.
Pause BM, Krauel K, Sokja B, Ferstl R (
1999
) Is odor processing related to oral breathing?
Int J Psychophysiol
 
32
:
251
–260.
Rausch R, Serafetinides EA (
1975
) Specific alterations of olfactory function in humans with temporal lobe lesions.
Nature
 
255
:
557
–558.
Rausch R, Serafetinides EA (1975b) Human temporal lobe and olfaction. In: Olfaction and taste V (Denton DA, Coghlan JP, eds), pp. 321–324. New York: Academic Press.
Rausch R, Serafetinides EA, Crandall PH (
1977
) Olfactory memory in patients with anterior temporal lobectomy.
Cortex
 
13
:
445
–452.
Royet JP, Koenig O, Gregoire MC, Cinotti L, Lavenne F, Le Bars D, Costes N, Vigouroux M, Farget V, Sicard G, Holley A, Mauguière F, Comar D, Froment JC (
1999
) Functional anatomy of perceptual and semantic processing for odors.
J Cogn Neurosci
 
11
:
94
–109.
Royet JP, Hudry J, Vigouroux M (2000a) Application de l'Imagerie Cérébrale à l'Etude de l'Olfaction. In: Rencontres IPSEN en ORL (Christen Y, Collet L, eds), pp. 85–99. Neuilly-sur-Seine: Editions Irvinn.
Royet JP, Zald D, Versace R, Costes N, Lavenne F, Koenig O, Gervais R (
2000
) Emotional responses to pleasant and unpleasant olfactory, visual, and auditory stimuli: a positron emission tomography study.
J Neurosci
 
20
:
7752
–7759.
Sakuma K, Kakigi R, Kaneoke Y, Hoshiyama M, Koyama S, Nagata O, Takeshima Y, Ito Y, Nakashima K (
1997
) Odorant evoked magnetic fields in human.
Neurosci Res
 
27
:
115
–122.
Savic I, Bookheimer SY, Frien I, Engel J (
1997
) Olfactory bedside tests: a simple approach to identify temporo-orbitofrontal dysfunction.
Arch Neurol
 
54
:
162
–168.
Seeck M, Mainwaring R, Ives J, Blume H, Dubuisson D, Cosgrove R, Mesulam MM, Schomer DL (
1993
) Differential neural activity in the human temporal lobe evoked by faces of family members and friends.
Ann Neurol
 
34
:
369
–372.
Seeck M, Schomer DL, Mainwaring R, Ives J, Dubuisson D, Blume H, Cosgrove R, Ransil BJ, Mesulam MM (
1995
) Selectively distributed processing of visual object recognition in the temporal and frontal lobes of the human brain.
Ann Neurol
 
37
:
538
–545.
Shipley M, Reyes P (1991) Anatomy of the human olfactory bulb and central olfactory pathways. In: The human sense of smell (Lain DG, Doty RL, Breipohl W, eds), pp. 29–60. Berlin: Springer Verlag.
Silver WL, Moulton DG (
1982
) Chemosensitivity of rat nasal trigeminal receptors.
Physiol Behav
 
28
:
927
–931.
Silver WL, Masson JR, Marshall DA, Maruniak JA (
1985
) Rat trigeminal, olfactory and taste responses after capsaicin desensitization.
Brain Res
 
333
:
45
–54.
Stuiver M (1958) Biophysics of the sense of smell. PhD thesis, University of Gröningen.
Swanson LA, Petrovich GD (
1998
) What is the amygdala?
Trends Neurosci
 
21
:
323
–331.
Talairach J, Bancaud J (
1973
) Stereotaxic approach to epilepsy methodology of anatomo-functional stereotaxic investigations.
Prog Neurol Surg
 
5
:
297
–354.
Talairach J, Tournoux P (1988) Co-planar stereotaxic atlas of the human brain: 3-dimensional proportional system: an approach to cerebral imaging. New York: Thieme Medical.
Truex RC, Carpenter MB (1969) Human neuroanatomy. Baltimore, MD: Williams Wilkins.
Winer BJ (1962) Statistical principles in experimental design. New York: McGraw-Hill.
Zald DH, Pardo JV (
1997
) Emotion, olfaction, and the human amygdala: amygdala activation during aversive olfactory stimulation.
Proc Nat Acad Sci USA
 
94
:
4119
–24.
Zald DH, Donndelinger MJ, Pardo JV (
1998
) Elucidating dynamic brain interactions with across-subjects correlational analyses of positron emission tomographic data: the functional connectivity of the amygdala and orbitofrontal cortex during olfactory tasks.
J Cereb Blood Flow Metab
 
18
:
896
–905.
Zatorre RJ, Jones-Gotman M (
1991
) Human olfactory discrimination after unilateral frontal or temporal lobectomy.
Brain
 
114
:
71
–84.