lhermitte's peduncular hallucinosis

an historical perspective

"Lhermitte's Peduncular Hallucinosis" is a paper by Anthony H. Risser and Frank C. Powell, which was presented at the 45th Annual Meeting of the American Academy of Neurology in New York, New York (April 1993). If cited, please use that citation.









In 1922, the French neurologist Jean Lhermitte described a patient with bizarre visual hallucinations associated with clinical signs suggesting focal midbrain and pontine involvement. A few years later, van Bogaert described similar features, naming the phenomenon "peduncular hallucinosis." At that time, lively debate about the origin and nature of hallucinations per se reflected both psychoanalytic concerns and neurological localization concerns. Lhermitte's cases and those of others, such as de Morsier and van Bogaert, presented a form of visual hallucination not localizable to central nervous system regions of direct importance to visual processing and analysis. Lhermitte's work and those initial case presentations localizing the lesion to midbrain and pontine involvement had impact upon the debate about hallucinations at that time. The advent of computed tomography and, then, of magnetic resonance imaging have confirmed the localization of these unusual hallucinations and additional case reports have confirmed their basic phenomenology (e.g., they do not invoke fear or reflect altered reality testing, but may evoke the interest or amusement of patients; being described by vivid and bright coloring). This paper describes the initial cases and explains the historical refinements in both the localization and the phenomenology of peduncular hallucinosis as we know it today.


Hallucinations have generated interest central to the development of various aspects of diagnostic thought in psychiatry, dating from Esquirol's description in 1837. They possessed an admittedly less central, though still important, interest to clinicians and localization theorists in neurology beginning with Hughlings Jackson and continuing into 20th-century neurology. Psychoanalytic and behavioral concepts developed and existed on the one hand, while appreciation for hallucinatory phenomena occurring in central disease processes and peripheral (sensory) disease existed on the other.

Not readily understandable from either perspective, however, was the 1922 case report by the French neurologist Jean Lhermitte (1877-1959). Lhermitte, a neurologist with a sustained interest in what is variably described today as neuropsychiatry or organic psychiatry, had considerable interest in the midbrain, the thalamus, and the manifestations of their disordered functioning. In the view of some, his work on hallucinatory phenomena and that on body-schema disorders provided exemplars for the interplay of neurological and psychiatric knowledge at the time (e.g., Lhermitte, 1951a, on autoscopia; Riese, 1960).

Lhermitte's 1922 case report described a 72-year-old female who experienced bizarre visual hallucinations. These hallucinations involved common animals (e.g., chickens) possessed with a strange appearance, people attired in costume, and children playing. Although the patient knew the hallucinations were not real, she would sometimes attempt to touch them. The images occurred late in the day, particularly at dusk. This patient's hallucinations were observed in association with signs suggesting a focal infarct involving the midbrain and pons (viz., left ptosis, total external ophthalmoplegia, lateral pulsion when walking, right dysmetria, intention tremor, and a Babinski sign). Lhermitte contended that the lesion accounted for the visual hallucinations as well. The patient was not followed to autopsy. The hallucinations were also associated with nocturnal insomnia and daytime somnolence, features important to Lhermitte in presenting the hypothesis that the hallucinations reflected a consequence of derangement in the sleep-wake cycle.

Perhaps the clearest statement at that time regarding the nature of hallucinations in neurological disease was a contemporary paper by Horrax in 1923. Horrax presented the literature to date, beginning with an 1879 report by Westphal and including cases from individuals such as Hughlings Jackson. Visual hallucinations that were unformed or elemental (e.g., lights or spots) were associated specifically with the calcarine area or, more broadly, with the occipital lobe. He also presented reports of his own patients with temporal lobe tumors, examined in Cushing's neurosurgical clinic. The majority of these patients, if they hallucinated, reported formed hallucinations (e.g., objects or people). Horrax and Putnam continued their work with a 1932 paper examining visual hallucinations and visual field disturbances in patients with tumors limited to the occipital lobe. Thus, there existed a body of clinical evidence suggesting that visual hallucinations could have localizing value depending upon their nature (unformed vs. formed).

Nevertheless, midbrain and pontine lesions (such as that proposed by Lhermitte) and the reported formed visual hallucinations associated with them did not easily figure into that schema. Lhermitte's case presented a form of visual hallucination not localizable to a central nervous system region of known direct importance to visual processing and analysis. In 1924, Ludo van Bogaert reported the case of a 59-year-old female with a history of rheumatic heart disease who developed vertigo, diplopia, and ataxia. Clinical findings were right IIIrd nerve palsy, left-sided dysmetria and hyperreflexia, and a gait ataxia, suggesting to van Bogaert a lesion of the red nucleus. At evening time, she would see the head of a dog or an image of an animal's body projected onto a wall of her home. The images were notable for their colors, specifically the colors red and green. He followed the case to autopsy a little over a year later, providing the first pathological verification. An extensive paramedian midbrain infarction involving the cerebral peduncle, medial substantia nigra, red nucleus, superior cerebellar peduncle, IIIrd nerve fibers, posterior longitudinal fasciculus, and periaqueductal gray was noted. He proposed, in 1927, to use the term "l'hallucinose pédonculaire" (peduncular hallucinosis) to describe the phenomenon. As has been subsequently accepted, the use of the term "pédonculaire" was not employed as a term limited to the cerebral peduncles, but to refer to the midbrain. With this case, van Bogaert provided the initial autopsy verification of the midbrain localization of the lesion presumed to result in these visual hallucinations.

Disagreement existed between Lhermitte and van Bogaert as to an explanation of the hallucinatory phenomenon. Lhermitte considered the condition to be a dissociated state of sleep in which the visual hallucinations were dreams occurring in a state of relative wakefulness. van Bogaert, while agreeing with Lhermitte's clinical judgment on the proposed localization, did not believe that the hallucinosis was a feature of deranged sleep-wake states. His case, specifically, apparently did not experience a sleep disorder, although the hallucinations only occurred in the evening. He proposed that the hallucinations were a state of ego dissolution with loss of ability to distinguish external reality from imagination. In this manner, some of the more-traditional psychoanalytic concepts entered into his interpretation. This interpretation was later decried by de Morsier (1935). de Morsier felt that van Bogaert was forgetting the lesional basis as "he lowers himself" to a general psychiatric level (i.e., a breakdown of ego functioning) to attempt to explain the phenomenon.

Lhermitte went on the present a number of additional cases (e.g., Lhermitte, 1932; Lhermitte, 1949; Lhermitte, Levy, & Trelles, 1932). Lhermitte, Levy, and Trelles (1932), for example, described a man who developed visual hallucinations during a barbiturate-induced delirium. They disappeared several weeks prior to his death. On autopsy, they described "metachromatic granules" and "pigmentary degeneration" of neurons of the oculomotor nucleus, pontine reticular formation, and midline raphe. McKee et al. (1990) have called these changes of "dubious significance" and raise the possibility of the changes being "an artifact resulting from celloidin embedding."

de Morsier extended the proposed localization to include the thalamus. He did so in presenting several of his own cases, as well as presenting a restatement of the pathology of van Bogaert's reported case as extending rostrally to include the pulvinar. One case of his was of a woman with visual hallucinations of landscapes and people moving in file in front of her. She presented with clinical evidence suggesting a thalamic syndrome manifested by hemianesthesia. Another case (de Morsier, 1936) was of a 54-year-old woman with left-sided hemiparesis, a slight right-sided ptosis that diminished over time, left-sided hypotonia, upward and right-sided gaze difficulties, and some dysarthria. She experienced visions of people she knew and "fantastic animals" such as a beautiful pheasant appeared to her "in a natural way 'like a cinema.'" Sometimes they appeared a brilliant white in color, almost fluorescent in its nature. Her visions moved from left to midline and then disappeared. They occurred at night-time as she was falling sleep but, over time, they would occur upon awakening and during the day-time. Insomnia had been present, but de Morsier argued that there was no temporal association between the time when she experienced sleep problems and the period during which she experienced the hallucinations. He argued that a right cerebral peduncular vascular lesion was present and responsible for the hallucinosis. de Morsier presented an essentially visually based localizational explanation of peduncular hallucinations by considering that the phenomenon was due to damage to retrochiasmic optical pathways.

Lhermitte's work and those initial case presentations localizing the lesion to midbrain and pontine regions had an impact upon the debate about hallucinations at that time. While de Morsier, in implicating the thalamus, returned peduncular hallucinosis to pathway connections of the brain with clearer relations to visual functioning, Lhermitte's basic explanation placed the phenomenon in the context of derangements in the sleep/wake cycle. In doing so, he preceded current explanations invoking involvement of the ascending reticular activating system and its impact on sleep stages.


Beyond the initial case reports of Lhermitte, de Morsier, and van Bogaert, subsequent descriptions of peduncular hallucinations have not presented any significantly different perceptions about the phenomenon. However, the elaboration of additional cases did allow for the understanding that different etiologies other than those with an acute cerebrovascular basis might result in the hallucinations. Analysis of the newer cases also allowed for a better understanding of localization (reported below in Part 3) and for a comparison of the clinical phenomena (Part 4).

In 1951, Lhermitte's (1951b) French monograph on hallucinations, "Les hallucinations: Clinique et physiopathologie," was published. The monograph covered the spectrum of hallucinatory phenomena and the fairly substantial number of his clinical contributions, but the primary chapter on visual hallucination emphasized peduncular hallucinosis and presented 5 cases.

Rozansky (1952) noted these hallucinations as a potential complication of vertebral angiography. Caplan's (1980) presentation on the various manifestations on the "top of the basilar" syndrome reacquainted clinicians with the phenomenon of peduncular hallucinosis as a rare feature of rostral brainstem infarction. Although, in the context of his broad review, he did not discuss these cases in detail, he noted a patient of his with a unilateral midbrain and thalamic infarct who would see "a colored parrot with beautiful plumage" on his right. He mentioned a second patient with "episodic posterior hemispheral ischemia" who would awaken in the middle of the night to see "pictures of his grandmother flashed on the wall to his left, as if projected in a home movie" and also experienced auditory hallucinations of a knocking noise, "as if rocks were in a car engine." He noted the presence of sleep abnormalities in his patients with supposed peduncular hallucinosis and reflected that the pathological anatomy was probably not limited to the midbrain alone. Another 1980 case reported peduncular hallucinosis in the context of circulatory insufficiency due to vertebral artery hypoplasia (Bossi, Nobili, & Benedetto, 1980).

As will be reported below, the availability of CT and MRI aided in the identification of brainstem and thalamic disease processes. Regarding peduncular hallucinosis, this added to the diversity of anatomic localization for proposed cases, particularly over the past several years. Geller and Bellur (1987) reported the first in vivo anatomic confirmation of a mesencephalic lesion causing peduncular hallucinosis using MRI.

Dunn and colleagues (1983) reported not only the first pediatric case (in a 6-year-old male), but also the first presentation resulting from brainstem compression, secondary to cystic craniopharyngioma formation. Hallucinations resolved following decompression with cyst drainage. Harada and colleagues (1991) reported peduncular hallucinosis in a 65-year-old female beginning within several days after an aneurysmal subarachnoid hemorrhage caused by a basilar-superior cerebellar artery rupture. They suggested that the hallucinations were a consequence of vasospasm of the perforating arteries supplying the upper brainstem.

McKee and colleagues (1990), with an autopsy study of an 83-year-old male with peduncular hallucinosis, showed the presence of a bilateral and symmetric infarct limited to the substantia nigra pars reticulata.


It is clear that the original clinicians were well aware that the causative lesion was not strictly circumscribed to the cerebral peduncles, but did include other areas of the mesencephalon. de Morsier further considered an important facet of lesion localization to be its extension to include the thalamus.

Subsequent cases, if anything, have allowed for different etiologies and anatomic regions to be considered as causative of peduncular hallucinosis. Yet, as Geller and Bellur's (1987) MRI study indicated, isolated mesencephalic lesions are a sufficient cause for the presentation of the hallucinosis.

MRI has been especially useful in documenting lesions in patients with peduncular hallucinosis. However, its use has not redefined the proposed anatomical localization. First, Geller and Bellur (1987), as noted above, used MRI to document the traditional localization of the lesion to the mesencephalon. Next, however, Feinberg and Rapcsak (1989), used MRI to document a case without any mesencephalic involvement, following paramedian thalamic infarction in an 83-year-old male. Finally, Kolmel (1991) reported MRI evidence of bilateral ischemic lesions of both the mesencephalon and the thalamus in a 56-year-old male with hallucinations.

Hattorri and colleagues (1988) reported that MRI revealed lesions in the right putamen and in the pons near the junction of the basis and the tegmentum in a 73-year-old male with the opsoclonus-myoclonus syndrome and accompanying visual hallucinations of a big rock or of a toy monster moving in front of him, described as being "vivid, colourful, and non-threatening." Serra Catafau, Rubio, and Peres Serra (1992) reported MRI findings of a right posterior thalamic infarct in the case of a 68-year-old male who developed vivid visual hallucinations (i.e., lions jumping into his room through a window, men's heads) associated with left hemiparesis and left-sided parasthesias. The following regions were involved: the lateral portion of the pulvinar, lateral geniculate body, posterior portion of the ventral nuclei, reticular nucleus, as well as a portion of the posterior limb of the internal capsule. The authors suggested that the reticular nucleus involvement was particularly important for the hallucinations.


[4-1] DISTINGUISHING PEDUNCULAR HALLUCINOSIS FROM OTHER VISUAL HALLUCINATIONS. Visual hallucinations occur in a variety of psychiatric, neurologic, ophthalmologic, and systemic toxico-metabolic disorders. Hallucinations with a neurologic etiology need to be distinguished as to whether they are related to seizure activity or not, as irritative hallucinations tend to be stereotypic in nature and in form, may have localizing significance, and may be identified as part of a patient's aura.

Those hallucinations that are features of the psychotic disorders are perhaps the most commonly appreciated hallucinatory phenomena. They may be observed in the schizophrenic disorders, mood disorders, and also may be a feature in conversion disorder. Diagnostically, hallucinations need to be differentiated from delusions, confabulations (e.g., Cummings & Miller, 1987) and, if observed in the context of delirium, from illusions and misperceptions (e.g., Lipowski, 1990).

In addition to the clinical distinction as to the presence of a seizure disorder, hallucinations have generally been characterized as being either "irritative" or "release" phenomena (Cogan, 1973). Peduncular hallucinosis is considered to be a "release" phenomenon, independent of seizure activity, although one case (McKee et al, 1990) reported the onset of hallucinations after the occurrence of a tonic-clonic convulsion. As a "release" phenomenon, peduncular hallucinosis is in the same class of visual hallucination as hypnagogic hallucinations and those observed in sensory-deprivation states.

Hemispheric lesions may result in visual hallucinatory perceptions. The historical distinction between elemental hallucinations associated with occipital lobe involvement and formed or more-complex ones being associated with temporal lobe lesions remains, but has been challenged (Weinberg & Grant, 1940). Hécaen and Angelergues (1965, as reported in Hécaen & Albert, 1978) reported the frequency of various hallucinations in a relatively large series of patients with documented unilateral retrorolandic tumors. They reported a higher frequency of complex visual hallucinations in patients with right-hemisphere tumors. Recently, Starkstein and colleagues (1992) have re-visited this literature to again suggest that a lateralization effect exists, with hallucinations being observed more frequently in patients with right-hemisphere lesions.

Of hemispheric lesions causing hallucinations, most important in terms of a differential comparison to peduncular hallucinosis are those lesions that cause visual field defects. Complex visual hallucinations with most of the features of peduncular hallucinosis may be observed to occur in the disrupted visual fields of a significant number of patients with field cuts, the only difference perhaps being whether or not the hallucinations are limited to visual field sectors (e.g., Kolmel, 1985; Lance, 1976).

The most common ophthalmologic-caused hallucinatory syndrome is the Charles Bonnet syndrome (e.g., Rosenbaum et al., 1987). First described in 1769 by Charles Bonnet, these visual hallucinations are associated with decreased visual acuity in the elderly and, like those of peduncular hallucinosis, are visually vivid, formed and complex in nature, and rarely evoke fear in individuals experiencing them.

Optic nerve disorders and migraine may result in hallucinatory perceptions. Toxico-metabolic derangements may result in hallucinatory experiences that often evoke fear in patients, the typical example being the visual and tactile hallucinations of Delirium Tremens (Lipowski, 1990).

While some argue that hallucinations and hallucinosis reflect a common underlying phenomenology that varies in its manifestation with the clinical context, others (e.g., Starkstein et al., 1992) argue that a distinction between the two be made on the basis of whether the patient believes that the non-existent sensory perception is real (an hallucination) or unreal (hallucinosis). Using this distinction, some have preferred to employ the term "pseudohallucinations" for those false perceptions that are recognized as being unreal. Kolmel (1991), however, takes a somewhat opposite point of view, arguing that peduncular "hallucinations" must be distinguished from hallucinosis, which to him are long-lasting phenomena without neurological deficit. Additionally, he argues that they are not "pseudohallucinations." Lhermitte himself appeared to have little use for any distinction between "true" and "pseudo" hallucinations, stating:

"...it is an error to want to set up a barrier between the hallucinations which are questioned and admitted to be illusions of the senses, or a hallucinosis, and, on the other hand, the hallucinations which are not criticized and which would be the only real hallucinations for some authors (Lhermitte, 1951a, pg. 432)."