OUP user menu

Localizing value of epileptic visual auras

Christian G. Bien, Felix O. Benninger, Horst Urbach, Johannes Schramm, Martin Kurthen, Christian E. Elger
DOI: http://dx.doi.org/10.1093/brain/123.2.244 244-253 First published online: 1 February 2000


It is difficult to differentiate between seizures of occipital or temporal lobe origin in patients with focal epileptic seizures associated with visual aura. These are often suspected to originate from the visual cortex, which causes reluctance to propose resection as treatment for the affected patient. To determine the value of localizing different types of visual auras, we report on 20 patients experiencing visual aura from a series of 878 surgically treated patients suffering from intractable focal seizures. In all of these patients, a morphological abnormality was identified on MRI (n = 18) or cranial CT (n = 2). These abnormalities were shown to represent the morphological correlate of the epileptogenic zone in each case, as demonstrated by intracranial ictal EEG recordings and/or seizure freedom after focal resective surgery. Elementary hallucinations, illusions and visual loss were reported not only by all patients with occipital lobe epilepsy, but also by patients with occipitotemporal and anteromedial temporal seizure onset. Complex hallucinations never occurred in occipital lobe seizures but were present in the two other groups. The same correlation was found for concentric changes of visual field (tunnel vision), a newly described ictal phenomenon. We conclude that elementary hallucinations, illusions and visual loss, although typical for occipital lobe epilepsy, can also occur in anteromedial temporal or occipitotemporal seizures and are therefore not a discordant feature in presurgical evaluation of patients with suspected temporal lobe epilepsy. Complex hallucinations and tunnel vision, however, should be considered concordant only with the assumption of an anteromedial temporal or occipitotemporal seizure onset.

  • epilepsy
  • focus localization
  • visual auras
  • temporal lobe epilepsy
  • occipital lobe epilepsy


Visual auras of different kinds are well known in partial seizures. The possible significance of attributing those auras to the site of seizure onset, however, has not been extensively studied. The best known examples are visual phenomena such as elementary hallucinations or visual loss, which occur in occipital lobe epilepsy (Huott et al., 1974; Babb et al., 1981; Blume et al., 1991; Salanova et al., 1992; Williamson et al., 1992; Fried et al., 1995; Shahar et al., 1996; Kuzniecky et al., 1997; Aykut-Bingol et al., 1998). However, visual phenomena may also occur in temporal lobe seizures and erroneously lead to the assumption of an occipital origin. Besides complex visual hallucinations (Penfield and Jasper, 1954; Penfield, 1958; Penfield and Perot, 1963; Arseni and Petrovici, 1971; Halgren et al., 1978; Gloor et al., 1982; Wieser, 1982), some authors have mentioned simple, occipital-like visual experiences in temporal lobe seizures (Russell and Whitty, 1955; French et al., 1993; Fried et al., 1995). To evaluate whether different types of ictal visual phenomena are related to particular sites of seizure origin, we retrospectively analysed 20 patients from a database containing 878 patients who had been operated on for medically intractable seizures at our epilepsy centre, in which sufficient data concerning visual phenomena and seizure onset zone had been obtained.



From the database of 878 patients operated on for medically intractable seizures at our institution between January 1988 and December 1998, we identified 39 patients who had experienced ictal visual phenomena. Only the 20 patients in which the epileptogenic focus had been determined with sufficient certainty were included in this study. The criteria for focus localization were as follows: (i) focus localized by intracranial recording of typical seizures (n = 8); and (ii) complete seizure freedom after resection of the suspected epileptogenic area (n = 17). Five patients fulfilled both criteria. The remaining 19 patients were excluded due to unclear focus localization (patients who were not seizure free postoperatively were operated on at different locations or were treated by procedures other than focal resections, e.g. hemispherectomy or callosotomy). The patients' ages at the time of evaluation ranged from 5 to 42, mean 24.5 ± 9.8 years. Follow-up continued for a mean of 22.2 ± 19.1 months (range 3–72 months). Data concerning the patients' history, clinical presentation, and results of non-invasive and invasive presurgical evaluation as well as histological diagnosis are given in Table 1.

View this table:
Table 1

Presurgical data, surgical procedure and outcome of patients with visual auras

Patient no., initialsSexAge at surgery (years)Initial seizure featuresNeuro-logical deficitLesion on MRI (CCT)Interictal surface EEGSPECTPETSurgical procedurePathologyFollow-up outcome
M = male; F = female; VA = visual aura; QA = quadrantanopia; LLQ = left lower quadrant; RLQ = right lower quadrant; RUQ = right upper quadrant; CCT = cranial computed tomography; MRI = magnetic resonance imaging; R = right; L = left; hem = hemispheric/hemisphere; temp = temporal; occ = occipital; front = frontal; par = parietal; centr = central; med = medial; bas = basal; lat = lateral; ant = anterior; post = posterior; gen = generalized; sw = sharp waves; ssw = sharp-slow-wave complexes; spw = spike-wave complexes; ict = ictal; hypo-p = hypoperfusion; hyper-p = hyperperfusion; hypo-m = hypometabolism; AMTL = anteromedial temporal lobe; MST = multiple subpial transsections; LE = lesionectomy; AHE = amygdalohippocampectomy; ECoG = intraoperative electrocorticography; HS = hippocampal sclerosis; sf = seizure free; m = month.
01, KOM29VAdysarthria, ataxiaCCT: L hem atrophy; L occ-med unclear lesiongen. sswL temp-med hypo-p; L temp-lat, R temp-med hyper-pMST L occ-par-temp15 m, not sf, no more VA
02, LKF30VAnoneMRI/CCT: calcified lesion L occ (infarct?)L temp swLE and MSTvascular malformation6 m, not sf
03, GSF30VAQA RLQCCT: L occ calcificationgen theta and swL temp hypo-pL occ resection under ECoGtuberous sclerosis72 m, sf
04, MPM24VAQA LLQL occ cystic lesionpar-occ spwL occ, L temp hypo-pL occ LEependymal cyst24 m, sf
05, PAM19VAnoneCCT: L occ cystic lesionL occ thetaL occ LE under ECoGposthaemorrhagic cyst60 m, sf
06, BSF16epigastric aura, then VAnoneR occ-temp cystic tumour (DNT?)L temp, R centr slowingict: R par, L temp hyper-pR temp-par-occ hypo-mR temp-occ LEganglioglioma I°12 m, sf
07, SOF13VAnoneR occ old haemorrhagegen spwR occ LEold hemorrhage10 m, sf
08, KTF5VAnoneR occ-temp hyperintense lesiongen thetaR occ hypo-pR occ LEganglioglioma I°18 m, sf
09, KCF19VAnoneL temp-post lesionL temp-occ swL temp and R par-occ hypo-pAMTL resection + LEastrocytoma II°40 m, sf
10, LRF16VAnonelesion below L trigonumbilateral sswbelow L trigonum hypo-mLEganglioglioma I°21 m, sf
11, KSF13VAnoneR parahippocampal gyrus: cystic lesionL hem>R hem thetaLE plus R hippocampusastrocytoma II°6 m, sf
12, WSF16VAnoneL temp-bas cystic lesionL temp swict: R temp hypo-p inter: L temp hypo-pL temp-par hypo-mLEastrocytoma II°6 m, sf
13, JaMF25VAQA RUQperiventricular heterotopia bilateralL temp swL>R temp hypo-pR hem hypo-mL temp-bas lesion resectionnodular cortical heterotopia12 m, sf
14, BWM42I: VA II:prickling R hand/arm III: staringparesis R armlesion L hippocampusL temp swL ant-temp hypo-mL AHEhamartia3 m, not sf, no more VA
15, JuMM30VAnoneL HSR/L temp swL temp hypo-mL AHEHS6 m, sf
16, RPF36VAnoneL hippocampal atrophyL temp thetaL AHEHS12 m, sf
17, SJM22VAnoneR HSR temp swR temp hypo-mR AHEHS42 m, sf
18, HCF38VAnoneL HSgen swL temp hypo-mL AHEHS12 m, sf
19, SMM29I: VA II: gustatory auranoneR HSR front-temp swR temp hypo-mR AHEHS30 m, sf
20, FWM37VAnoneR HSnormalR temp, L temp-pol, L par-occ hypo-pR temp hypo-mR AHEHS36 m, sf


Preoperative MRI was performed in 18 patients, and for the present study images were available in 15 cases. In the other three patients only postoperative MRIs were available. The investigations were performed on a 1.5 T scanner (Gyroscan ACS-II, Philips Medical Systems, Best, The Netherlands). Patients were studied by a standardized protocol for presurgical evaluation of patients with focal seizures, in which slices are routinely angulated along or perpendicular to the longitudinal axis of the hippocampus. In four patients, slices were angulated along the AC–PC (anterior commissure–posterior commissure) line. Two patients were studied preoperatively by cranial CT, because they were investigated prior to the introduction of routine MRI during the presurgical evaluation. MRI or cranial CT slices of each patient which most clearly showed the extent and/or topographic relationship of the lesion to important structures (e.g. hippocampus) were transformed into schematic sketches and are given in Table 2.

Assessment and classification of visual auras

Patients' reports on their seizure symptoms and signs, as reported on hospital admission for presurgical evaluation, were obtained from the hospital files. No established classification of visual auras was found in the literature. We used the following groups, which are most commonly distinguished from each other (Penfield and Jasper, 1954; Ludwig and Marsan, 1975; Gloor et al., 1982; Wieser, 1982; Sveinbjornsdottir and Duncan, 1993). (i) Elementary or simple hallucinations, i.e. perception of white or coloured, unformed or formed phenomena, the latter in various, but always simple forms and shapes such as a line, a square, spots, a zigzag of lines, flashes and so on, stationary or moving. (ii) Visual illusions, i.e. an alteration in perception of objects like micropia/macropia (dysmegalopsia), dyschromatopsia, achromatopsia, metamorphopsia or kinetopsia. (iii) Ictal blurred vision or amaurosis in the whole visual field or confined to a part of it (quadrantanopia, hemianopia). (iv) Complex visual hallucinations of animals, people or scenes, of numerals or letters, stationary or moving. As has been observed, in complex hallucinations much more `input' is given by the individual than in elementary hallucinations or illusions, e.g. elements from memory, unfulfilled wishes or intense affects related to them (Gloor et al., 1982).

Ictal EEG recordings

In nine patients, intracranial EEG recordings were obtained. We used subdural strip electrodes and bilateral intrahippocampal depth electrodes implanted stereotactically along the longitudinal axis of the hippocampus, in one case a depth electrode inserted into a heterotopia was used in addition (patient 13, JaM); see Table 3. The details of implantation procedures and intracranial EEG recordings with those electrodes have been described in earlier publications (Behrens et al., 1994; Dümpelmann and Elger, 1998; van Roost et al., 1998). In five patients, electrical stimulation via chronically implanted electrodes was performed for provocation of seizures. In the remaining 11 patients, continuous EEG recordings were performed using the international 10–20 system of surface electrodes plus sphenoidal electrodes.

View this table:
Table 3

Ictal EEG recordings

Patient no.InitialsElectrodesSeizure onsetLocalizing to lesionElectrical stimulation
sd = subdural electrodes; dp = depth electrodes; occ = occipital; temp = temporal; lat = lateral; med = medial; bas = basal; hem = hemispheric; hipp = hippocampus; par = parietal; max = maximum; LVFA = low voltage fast activity; + = localizing to lesion; (+) = localizing to lesion and additionally to neighbouring regions; VA = visual aura; sz = seizure; RVF =right visual field.
With intracranial EEG
01KOsd: R: occ-polar/-temp/-inter-hem; L: occ-lat/-temp/-inter-hem, temp-lat, centralmax L occ (beta)+L occ-polar: typical VA
02LKsd: L: occ-lat/-temp (Gyrus angularis); dp: L/R hippL occ-lat and occ-temp (LVFA)+not done
12WSsd: L/R temp-lat/-bas; dp: L/R hippL hipp, temp-bas (beta)+not done
18HCsd: L/R temp-lat/-bas; dp: L/R hippL hipp, temp-bas (alpha)+L temp-bas, L hipp: typical sz
20FWsd: L/R temp-lat/-bas; dp: L/R hippR hipp (alpha)+R hipp: typical VA
16RPsd: L temp-lat/-bas/-bas-occ R temp-lat/-bas; dp: L/R hippL temp-lat, temp-bas (alpha): subclinical sz(+)L/R hipp, L temp-lat, R temp-bas: typical VA
17SJsd: L temp-lat/-bas. R temp-lat/-bas/-bas-occ; dp: L/R hippR hipp (alpha)+no sz induction
13JaMsd: L temp-lat/-bas. R temp-lat, temp-bas ant/post, temp-bas-occ dp: L/R hipp, L intralesional (lat of inferior horn of lat ventricle)L temp-bas ant/post (LVFA)+L intralesional, L temp-bas-occ: elementary hallucinations RVFL temp-lat: metamorphopsia R hipp: complex hallucinations. No typical VA induced
14BWsd: L temp-lat/-bas, occ-lat, central; dp: L/R hippL hipp (alpha)+not done
With surface EEG
03GSL occ-par (theta)(+)
04MPL occ-central-temp (theta)(+)
05PAno sz recorded
06BSR temp-central (theta)(+)
07SOnot evaluable (artefacts)
08KTR occ (theta-delta)(+)
09KCL temp-med (theta)(+)
10LRno sz recorded
11KSR temp (theta-delta)+
15JuML temp (theta)+
19SMnot evaluable (artefacts)


Five patients had more than one visual aura phenomenon (04, MP; 05, PA; 10, LR; 14, BW; 15, JuM). A total of 27 different aura types were mentioned by the 20 patients. The localizations of the morphological correlates of epileptogenic areas were as follows: five patients (25%) had lesions confined to the occipital lobe. Another five patients (25%) were found to have abnormalities at the temporo-occipital junction or in the posterior part of the temporal lobe. Ten patients (50%) had pathological findings in anteromedial temporal structures (five of them had histologically confirmed hippocampal sclerosis); this region was equivalent to the part of the temporal lobe removed in standard anteromedial two-thirds resection. The aura types and the sites of MRI/cranial CT abnormalities are shown in Table 2; see also Table 4. Habitual clinical seizures were recorded in 17 patients (intracranial EEG, n = 8; surface EEG, n = 9). In the eight patients with implanted electrodes, the seizure origin could be localized to the site of the MRI/cranial CT abnormality; in the patients investigated by surface EEG seizure origin was not obtainable in two cases, the initial ictal activity localized to the site of the MRI/cranial CT abnormality in two cases and additionally to adjacent areas in five patients. In five patients visual auras could be induced by electrical stimulation via chronically stimulated implanted electrodes. For details on ictal EEG recordings and results of electrical stimulation, see Table 3.

View this table:
Table 4

Visual auras: types and origin

Elementary hallucinationsIllusionsVisual lossComplex hallucinationsConcentric changes (CC) of visual fieldTotal
MetomorphopsiaDyschromatopsiaTunnel visionScenic CC
Temp = temporal; occ = occipital; med = medial.
Occipital lobe4212 9
Temp-occ junction21111 6
Antero-med temp lobe21223212

Elementary hallucinations

Eight patients described elementary hallucinations, including flickering lights, silver-white stars, circles, flashes of light, and a blinking square. These phenomena appeared partly coloured, partly monochrome, and stationary or moving. Four of these patients belonged to the occipital lobe group (01, KO; 02, LK; 04, MP; 05, PA), two to the temporo-occipital group (08, KT; 10, LR), and two to the anteromediotemporal group (15, JuM; 14, BW). Six patients reported these elementary hallucinations as either being confined to or at least starting in the half of their visual field contralateral to the seizure focus. One patient with left temporomedial seizure onset (15, JuM) experienced the first ictal sign as tunnel vision (see below) and consecutive star-like flashes in the concentric lost part of his visual field, as pronounced on the right side. One patient (08, KT) did not identify a specific part of her visual field. In patient 13, JaM (bilateral periventricular heterotopia, seizure focus left anteromediotemporal as shown by intracranial recording of habitual seizures and complete seizure freedom after left temporobasal resection), electrical stimulation of the epileptogenic area and adjacent temporo-occipital contacts produced non-habitual elementary hallucinations (`coloured moving spots', `diffuse flickering', `circle') in the right visual field.

Visual illusions

Five patients experienced visual illusions: three described a distortion of proportions or contours or an increase or reduction in the size of objects of the real world, i.e. metamorphopsia (occipital group: 03, GS; 05, PA; temporo-occipital: 10, LR), two patients reported an intensified appearance of colours (dyschromatopsia, occipital: 04, MP; anteromediotemporal: 13, JaM). These phenomena affected either the whole visual field or only a part of it, and there was no differentiation between the groups according to the localization of their seizure onset. In patient 13, JaM (see above), non-habitual metamorphopsias (`ceiling is moving upwards', `things turn around a point in the middle of the visual field') could be elicited by electrical stimulation of the left temporolateral cortex.

Visual loss

Visual loss occurred as blurred vision, `white out' or scotoma. Such loss was reported by patients belonging to all three groups according to the localization of their lesion (occipital: 04, MP; 05, PA; temporo-occipital: 07, SO; anteromediotemporal: 14, BW; 20, FW). According to the patients' records, the visual loss could start in a specific part of their visual field (contralateral to the focus) or affect the visual field as a whole. There was no phenomenological criterion which differentiated between the three localizational groups.

Complex hallucinations

Three patients gave an account of complex hallucinations: moving heads (temporo-occipital: 09, KC), animals (anteromediotemporal: 19, SM) or little people (anteromediotemporal: 12, WS). These visual experiences appeared three-dimensionally, but elaborate scenes were not reported. In patient 13, JaM (see above), electrical stimulation of the right hippocampus led to a `reappearance of short dream sequences', which was not a habitual seizure element of this patient.

Concentric changes of visual field

Six patients described some form of ictal concentric change of their visual field. Four of them (temporo-occipital: 06, BS; anteromediotemporal: 11, KS; 15, JuM; 17, SJ) experienced a continuous progressive restriction of the visual field which did not lead to complete loss of vision, referred to as `tunnel vision'. The other two patients (anteromediotemporal, both hippocampal sclerosis: 16, RP; 18, HC) experienced this concentric change in a more scenic and non-stationary way that was associated with emotional distress: HC spoke about the frightening experience of an enlarging black ball approaching her. RP had the `visual feeling' as if being drawn into a hole or she saw herself standing at the edge of a volcano crater (dark, round). In this latter patient, who became seizure free after left-sided amygdalohippocampectomy, this type of visual aura could be elicited by electrical stimulation of either the hippocampus or of the temporolateral neocortex on either side, whereas no habitual seizure could be recorded during continuous intracranial monitoring.


Classification of visual aura types

Different types of ictal visual phenomena have not as yet been defined clearly. Most authors differentiate `simple' or `elementary hallucinations' from `complex hallucinations', where `simple hallucinations' are regarded as phenomena which are at most as elaborate as regular geometric figures and `complex hallucinations' are defined as visual appearances of things, faces or scenes. Since in most cases the patients recognize these phenomena as not really existing but can readily identify them as part of their epileptic sensations, one could more precisely call these phenomena pseudo-hallucinations. Since this term is not common in the literature on auras, we omit the prefix pseudo. Ictal illusions have been distinguished from hallucinations as in the former cases a real sensory input is misperceived by the patient during epileptic discharges (e.g. dyschromatopsia, metamorphopsia). Furthermore, patients in our study reported auras in which a concentric change of visual field was the predominant common feature. To our knowledge, this ictal phenomenon has not been described previously.

Localization of epileptogenic zones

The lesions demonstrated by MRI or cranial CT were shown to be the morphological correlates of the epileptogenic zone, by intracranial recording of spontaneous habitual seizures in eight patients and by surface recordings in another seven patients, which localized the focus by EEG to the region of the lesion. Seizure freedom was achieved in 17 patients by focal resective surgery. Every patient in our study was either seizure free after surgery or the focus could be determined by intracranial seizure recordings. For practical reasons with respect to the demands of presurgical evaluation we set up three groups of seizure onset areas: (i) an anteromedial temporal group; (ii) an occipital group; and (iii) an `occipitotemporal' group. Since there is no anatomical boundary between occipital and temporal lobe, all cases `in between', with lesions in the posterior part of the temporal lobe as well as those with lesions extending from one lobe to the other, were classified in this latter group. We found no patients with a seizure origin outside of these regions, e.g. in the parietal lobe. An assignment of the epileptogenic areas to histologically or functionally defined parts of the brain (e.g. Brodmann areas or V1, V2 and so on) was not attempted since such areas could not be determined in individual patients on the basis of the available data.

Elementary hallucinations, illusions and visual loss and occipital seizure origin

The correlation of certain types of visual auras with particular sites of seizure onset has been well documented: elementary hallucinations, visual illusion, partial or complete loss of sight evoked during spontaneous seizures (Huott et al., 1974; Babb et al., 1981; Blume et al., 1991; Salanova et al., 1992; Williamson et al., 1992; Fried et al., 1995; Shahar et al., 1996; Kuzniecky et al., 1997; Aykut-Bingol et al., 1998) or by electrical stimulation (Penfield and Jasper, 1954; Penfield and Perot, 1963; Wieser, 1982) have been described as occipital in origin. All of our five occipital patients had these kinds of ictal phenomena, which are obviously due to an ictal involvement of the primary visual cortex (elementary hallucinations, visual loss) or associated fields (illusions) (Penfield and Jasper, 1954).

Complex visual hallucinations and temporal lobe seizures

Complex hallucinations occur in spontaneous temporal lobe seizures as well as during temporal lobe electrical stimulation (Penfield and Jasper, 1954; Penfield, 1958; Penfield and Perot, 1963; Arseni and Petrovici, 1971; Halgren et al., 1978; Gloor et al., 1982; Wieser, 1982). Different results have been obtained depending on whether the complex hallucinations were associated with temporolateral/neocortical or temporomedial/limbic activation. Penfield and co-workers reported on complex hallucinations after stimulation of the temporal neocortex anteriorly as well as posteriorly primarily on the non dominant side (Penfield and Jasper, 1954; Penfield, 1958; Penfield and Perot, 1963). In contrast, Gloor and co-workers, as well as Wieser, found that complex visual phenomena did not occur unless there was an ictal involvement of limbic structures; they explained the discrepancy between the two results by the lack of depth of the electrode stimulation in the investigations performed by Penfield and his group (Gloor et al., 1982; Wieser, 1982). In our study, three patients reported complex hallucinations (09, KC; 12, WS; 19, SM), but only in patient WS was evaluation by intracranial electrodes carried out. Seizure onset was found to be temporomedial/basal (on MRI and on operation, a tumour was found in this region). In the other two patients, surface recordings were performed. The seizures in patient KC showed temporomedial (i.e. sphenoidal) onset on surface EEG. This patient had a temporoposteriorly located tumour that did not directly involve limbic structures, as judged by MRI and on surgical lesionectomy; in addition, anteromedial temporal lobe resection (including the hippocampus) was performed. Therefore, in her case it could not be ascertained whether the epileptogenic zone responsible for the complex hallucinations was located in neocortical or limbic structures. The third patient, SM, had hippocampal sclerosis and became seizure free after amygdalohippocampectomy. Even though in this case no evaluable seizure recording is available, the other preoperative findings unequivocally indicated the syndrome of unilateral medial temporal lobe epilepsy with hippocampal sclerosis, and since the patient became completely seizure free after right-sided amygdalohippocampectomy, it can be concluded that the epileptic activity arose from limbic structures. Therefore, two of the three cases with complex hallucinations support Gloor's and Wieser's assumption of a limbic origin of complex visual phenomena. On the other hand, we have no evidence in accordance with Penfield's hypothesis of a neocortical origin.

Elementary hallucinations, illusions and visual loss in anteromedial temporal lobe patients

Four patients in the anteromedial temporal lobe group (13, JaM; 14, BW; 15, JuM; 20, FW) reported ictal visual phenomena which would be considered by most epileptologists to be of occipital origin. One patient (14, BW) reported two such aura types, bringing the number of them in Table 4 to five. Elementary hallucinations in temporal (lateral or medial) epilepsy have occasionally been reported (French et al., 1993; Fried et al., 1995). In cases of post-traumatic epilepsy of probable temporal lobe origin, elementary hallucinations and ictal visual loss have been described. However, neither invasive EEG nor modern neuroimaging techniques were available in this study (Russell and Whitty, 1955). Gloor and associates, using temporal depth electrodes, found visual illusions in spontaneous seizures when limbic structures were involved, and they could elicit elementary hallucinations by electrical stimulation of limbic structures (Gloor et al., 1982). Penfield and Jasper found visual illusions by temporoposterior electrical stimulation (Gloor et al., 1982). Other studies using electrical stimulation have reported similar visual phenomena (Halgren et al., 1978; Gloor et al., 1982). Cases of occipital-like auras evoked by primary temporal lobe rhythmic activity have been explained by spread of epileptic activity—probably via the optic tract or the optic radiation—to the calcarine sulcus (Russell and Whitty, 1955; Gloor et al., 1982). In contrast, Halgren and co-workers related these phenomena to the intrinsic visual properties of temporomedial stuctures (Halgren et al., 1978). In our study, seizures of two patients from the anteromedial temporal group experiencing simple visual auras were recorded by intracranial EEG. In patient JaM, spontaneous seizures originated temporobasally (visual aura: dyschromatopsia); no rhythmic activity was recorded during these seizures from the most posterior (i.e. left occipital) electrode contact. Several types of simple visual auras (even types the patient did not experience during spontaneous seizures) were evoked by electrical stimulation of either temporal lobe. In these cases also, no rhythmic discharge over the (left) occipital lobe was registered. This, however, does not exclude the possibility of epileptic discharges in deeper parts of the calcarine sulcus propagated from the temporomedial focus. Patient FW had three spontaneous visual auras (visual loss) during seizures originating from his right hippocampus after intracranial electrodes had been implanted; the same type of aura was elicited several times by electrical stimulation of the same area. No occipital electrodes were implanted in this patient. From the results obtained so far, we can neither prove nor reject the above mentioned hypotheses.

Occipitotemporal junction group

The broad spectrum of visual aura types in the occipitotemporal group, which has also been observed by other investigators (Wieser, 1982; Bancaud, 1987; Palmini et al., 1993), is probably due to different patterns of seizure spread either to the visual cortex or to the anteromedial parts of the temporal lobe, or it may reflect intrinsic visual properties of the posterior temporal cortex. Since none of our patients in this group was studied using intracranial electrodes, no further electrophysiological data are available from the present study.

Complex hallucinations in occipital lobe epilepsy

Some authors have described patients with spontaneous seizures, which they considered to be of occipital origin, as having complex visual hallucinations; all these aura phenomena were considered to represent seizure spread to the temporal lobe, which was consistent with the precise descriptions of initial occipital signs and symptoms followed by complex hallucinations and other temporal lobe features (Salanova et al., 1992; Williamson et al., 1992; Aykut-Bingol et al., 1998). The phenomenon of occipital lobe seizures spreading into anterior parts of the brain, primarily temporolimbic structures, has been well described on the basis of depth electrode recordings (Babb et al., 1981; Olivier et al., 1982; Salanova et al., 1992; Williamson et al., 1992; Palmini et al., 1993). Penfield and co-workers could not elicit complex visual hallucinations from the occipital lobes by electrical stimulation (Gloor et al., 1982).

Concentric changes of visual field as ictal sign

Six patients reported progressive concentric changes of their visual fields as part of their habitual seizures. To our knowledge, this ictal phenomenon has not been reported in the literature so far. It could occur as mere constriction of visual field, mostly referred to as `tunnel vision', without particular emotional rapport. Two patients, however, experienced concentric changes of their visual field with scenic elements of frightening or dramatic character. Five of the six patients belonged to the anteromedial temporal group; four of them had hippocampal sclerosis. The one patient in the occipitotemporal junction group (06, BS) experienced epigastric discomfort as the first seizure sign, which was followed by tunnel vision. The close association of this kind of visual aura with a typical limbic seizure element speaks in favour of a common origin of both features. In one patient (18, HC) a typical aura was documented by intracranial EEG; seizure onset was recorded in the left hippocampal and temporobasal electrodes. By electrical stimulation, typical auras could be elicited from the hippocampal and the temporobasal electrodes on the focal side. The other patient with this aura type was studied by intracranial EEG (16, RP) but did not experience a full-blown seizure during continuous monitoring. By electrical stimulation concentric changes of her visual field could be evoked from either hippocampus, from temporolateral contacts on the focal side and from temporobasal electrodes on the contralateral side.

Concentric changes of visual field cannot be explained as epileptic involvement of the primary visual cortex. None of our occipital patients reported tunnel vision or similar phenomena. The emotional aspects and the scenic character in some of these auras, as well as the results of intracranial recordings and electrical stimulation, make a limbic origin (amygdala, hippocampus) most probable.


Phenomena which are typical for seizures of occipital lobe origin (elementary hallucinations, illusions like metamorphopsia or dyschromatopsia, visual loss) can also be experienced in temporally (anteromedially or posteriorly) originating seizures. Complex hallucinations or concentric changes of visual field, most probably of temporolimbic origin, were only found in patients with anteromedial or occipitotemporal foci. Patients undergoing presurgical evaluation with these kinds of visual auras as initial ictal phenomenon and suspected occipital lobe epilepsy should therefore be investigated by intracranial EEG to rule out a temporal seizure onset. In addition, functional imaging like PET (Henry and Chugani, 1998) or SPECT (Ho et al., 1995; Newton et al., 1995), or even magnetoencephalography (Ebersole, 1998), may be helpful to determine the epileptogenic focus in these patients.


This study was supported by Deutsche Forschungsgemeinschaft (EL 122/6-1, SFB 400) and BONFOR (research support programme of the University of Bonn, No. 111/28).


View Abstract