Purpose: Our goal was to use functional magnetic resonance imaging (fMRI) to demonstrate brain activation in patients with unirhinal, episodic, recurrent phantosmia who induced their phantosmia by coughing, sneezing, laughing or vigorous nasal inhalation and expiration, and inhibited it by sleep or performance of a Valsalva type maneuver.
Methods: Three patients with unirhinal phantosmia without change in taste or smell acuity were studied by fast low angle shot (FLASH) MRI and by echo planar imaging (EPI). Brain activation was measured following memory of two tastants (salt, sweet), memory of two odorants (banana and peppermint), actual smell of three odors (amyl acetate, menthone, pyridine), memory of phantosmia (and phantageusia, where applicable), phantosmia initiated spontaneously or by vigorous nasal inhalation and exhalation, phantosmia after inhibition by Valsalva, and these stimuli before and after treatment with the neuroleptic thioridazine. Activation images were derived using correlation analysis and ratios of areas of brain activated to total brain areas were calculated. Total activated pixel cluster counts were also used to quantitative total and regional brain activation.
Results: Sensory-specific brain activation was present in each section in each patient following memory of tastants and odorants, actual smell of each odor and memory, and initiation of and inhibition of phantosmia. Activation to odor memory after phantosmia initiation was very robust, whereas after phantosmia inhibition it was similar to that in normal subjects. Brain activation to unirhinal phantosmia was bihemispheric, independent of whether it was left or right sided or patient handedness. While phantosmia memory (in the absence of initiated phantosmia) produced extremely robust brain activation, after initiation and inhibition of phantosmia apparent brain activation decreased. These changes need to be related to shifting state of baseline brain activation and should be interpreted to reflect increased rather than decreased brain activation over that of phantosmia memory alone. Treatment with thioridazine inhibited brain activation to all stimuli including phantosmia and phantageusia memory, as it did previously in patients with birhinal phantosmia.
Conclusions: 1) Unirhinal phantosmia can be demonstrated by brain fMRI as can birhinal phantosmia; 2) unirhinal phantosmia can be initiated and inhibited by physiological maneuvers reflected by changes in fMRI brain activation; 3) fMRI brain activation of unirhinal phantosmia is bihemispheric and independent of peripheral side of phantosmia or patient handedness; 4) anterior frontal brain region plays a significant role in both phantosmia initiation and inhibition as, to some extent, do temporal brain regions; 5) activation of brain GABAergic systems appears to play a role in inhibition of unirhinal phantosmia; and 6) unirhinal phantosmia, similar to birhinal phantosmia, may reflect a type of maladaptive brain plasticity similar to that hypothesized to be responsible for phantom limb pain.