Donepezil in the treatment of musical hallucinations


Satoshi Ukai, MD, PhD, Department of Neuropsychiatry, Wakayama Medical University, 811-1, Kimiidera, Wakayama 641–0012, Japan. Email:


Abstract  Musical hallucinations (MH) typically occur among elderly individuals and are associated with hearing impairment. The authors describe a patient with features of typical MH who was successfully treated with donepezil, a cholinesterase inhibitor, as a combination therapy and who has not shown any subsequent cognitive decline for approximately 5 years. The efficacy of donepezil in this patient indicates that age-dependent dysfunction of cholinergic neurons might be related to the development of MH.


Musical hallucinations (MH) are the clinical phenomenon of hearing tunes or melodies without any relevant relation to external stimuli.1–3 Berrios showed that they were considerably more common among females, and that advanced age, deafness, and brain disease in the non-dominant hemisphere were important factors in their development.1 Pasquini and Cole reviewed 32 patients of MH who were more than 65 years old, excluding patients with focal neurological diseases and drug toxicity.4 They classified these cases as idiopathic MH, and found that they were predominantly females with hearing impairment, and that one-third of them were depressed.

Neuroleptics, antidepressants, and anticonvulsants have often been used for the treatment of MH on clinical experience, but they were ineffective in most cases. Treatment has occasionally been effective only when the MH occurred in conjunction with a specific disease, for example, with the use of antidepressants for depression, neuroleptics for schizophrenia, and anticonvulsants for underlying seizure disorders.2,4

In this case report, the authors describe a successful treatment with donepezil, a cholinesterase inhibitor, of an elderly patient with hearing impairment, who had symptoms fitting the criteria of idiopathic MH, but did not show any cognitive decline. The authors obtained her oral consent to publish this paper after explaining its details to her.


The patient was an 82-year-old woman. In 1996, she suffered from sensorineural hearing loss in both ears, and her hearing impairment had slowly worsened thereafter. She began wearing a hearing aid in early 1999.

One day in December 1999, she suddenly began to experience MH (Fig. 1). Subsequently, she heard voices singing old familiar songs accompanied by musical instruments almost all day long. About 1 month later, her MH changed into monotonous, simple melodies without singing but they were still heard almost continuously.

Figure 1.

The clinical course of the patient with musical hallucinations. Musical hallucinations were evaluated using visual analog scale.

Approximately 2 months after onset, she told her otorhinologist about her MH, and was treated with etizolam. Then, the intensity of her MH was slightly reduced. At 4 days later, the authors consulted with her for the first time. Except for her MH, she showed no apparent psychiatric symptoms such as depression, psychosis, or cognitive decline. She had no signs suggesting disturbance of consciousness, attention, or concentration. Her sleep–wake cycle was normal. Her Mini-Mental State Examination score was 29/30. Routine laboratory data were within normal ranges. Magnetic resonance imaging of the brain showed diffuse mild cortical atrophy that was within normal limits for her age. Electroencephalogram disclosed that thefrequency of the background activities was 8.5 Hz without focal abnormalities, and was within normal limits.

The authors initiated pharmacotherapy with etizolam, nicergoline, and fluvoxamine. About 3 weeks later, carbamazepine was added. Approximately 1 week after the start of carbamazepine, the intensity of her MH decreased slightly but they were still heard almost continuously. She was subsequently prescribed various combinations of medications. These included etizolam, nicergoline, an antidepressant (either fluvoxamine, milnacipran, or paroxetine), and an anticonvulsant (either carbamazepine or valproic acid). While the antidepressants and anticonvulsants were changed one by one, none of the drug combinations substantially reduced her MH, but the duration and intensity of her MH had slowly and gradually diminished. This reduction appeared to be due to the natural course of her condition rather than to her treatment with the various drug combinations. Then, the MH were heard only in quiet environments and had lessened to 7/10 on a visual analog scale.

In March 2001, The authors explained to her a possibility that donepezil, which is approved only for treatment of Alzheimer's disease, might reduce her MH, and the authors obtained her oral consent for pharmacotherapy with donepezil. Donepezil 2.5 mg/day was added, and paroxetine was discontinued in order to avoid increasing the number of her medications. Within a few days, her MH had diminished. At 2 weeks later, donepezil was increased to 5 mg/day and her MH were soon diminished markedly. She only experienced MH when she focused her attention on them in quiet environments. The monotonous melodies of simple sounds changed to short, simple rhythmic sounds. Her score on the visual analog scale declined to 2/10. Over approximately 20 subsequent months, she was treated with a combination of etizolam, nicergoline, valproic acid, and donepezil.

In November 2002, she complained that she worried about her family and that her MH had increased, although she showed no symptoms suggesting depression. Fluvoxamine was restarted and valproic acid was decreased, and the MH diminished back to the previous level. At 3 months later, valproic acid was discontinued and no changes in her MH occurred. Subsequently, for approximately 14 months, her therapy consisted of a combination of etizolam, nicergoline, fluvoxamine, and donepezil.

In March 2004, fluvoxamine was discontinued, resulting in no changes in her MH. At 1 month later, she was very busy tending to her sick brother, and discontinued her regular visits and medication without consulting us. In the next 2 months, her brother and her son's mother-in-law died in spite of her provision of intensive care for them. Then, she visited us again and said that her psychiatric symptoms had not changed in the last 3 months while she was off her medication, but that she had felt very tired about 10 days after her son's mother-in-law's death. She showed depressed mood, and had poor appetite and sleep, indicating that she was depressed. She also complained that her MH had increased as she became depressed.

The authors restarted her pharmacotherapy with etizolam, nicergoline, milnacipran, and donepezil. About 1 week later, her MH declined to the level they had been. After a few months, she recovered from depression. Since then, she has been maintained on the combination of such medications and has shown no remarkable changes in her psychiatric symptoms.


To the authors' knowledge, this is the first case report describing the successful treatment of MH using donepezil. In this patient, donepezil remarkably reduced her MH as a combination therapy, while it is unknown whether donepezil can reduce her MH as a monotherapy. Although she successfully responded to donepezil, she has not shown any cognitive decline during the approximately 5-year follow up. While previous studies have indicated that aging of the brain is one of the major predisposing conditions for MH, the relationship between aging of the brain and the development of MH is unclear.2 The efficacy of donepezil in this patient suggests that age-dependent dysfunction of cholinergic neurons might be related to the development of MH.

MH are often compared to Charles Bonnet syndrome (CBS).5,6 Typically, patients with CBS are elderly individuals with preserved intellectual function and the condition is often associated with visual impairments. While CBS is often resistant to pharmacotherapy with neuroleptics, antidepressants, and anticonvulsants,5,6 the authors previously reported the effectiveness of pharmacotherapy with donepezil in a patient with features of typical CBS.7

In idiopathic MH and typical CBS, elderly individuals suffer from auditory or visual impairments, respectively, and hallucinations related to the corresponding sensory modalities can subsequently develop. It has been speculated that their common pathophysiological mechanisms involve ‘release phenomenon’.8 According to this theory, when sensory inputs, which usually inhibit the generation of percept or memory traces, decrease below a certain threshold, previously recorded perceptions are released into awareness through the disinhibition of the brain circuitry that represents them. However, the neural basis of this theory is still unknown. The response to donepezil in the authors' patients with MH and CBS might suggest that age-dependent dysfunction of the cholinergic neurons is one of the predisposing conditions for ‘release phenomenon’.

Because donepezil has fewer adverse effects than neuroleptics, antidepressants, and anticonvulsants, donepezil can be the most effective and safest choice for the treatment of MH when used either as a monotherapy or a combination therapy in the elderly. Clinical trials should be conducted to evaluate its efficacy in the treatment of MH.