Robert I. Henkin

Functional Mri of Human Olfaction

PURPOSE Our goal was to use functional MRI (fMRI) to measure brain activation in response to olfactory stimuli. METHOD fMRI brain scans were obtained in 17 normal subjects (9 men, 8 women) using-multislice FLASH MRI in response to three olfactory stimuli (pyridine, menthone, amyl acetate) in three coronal brain sections selected from anterior to posterior temporal brain regions. Activation images were derived using correlation analysis, and ratios of areas of brain activated to total brain areas were calculated. RESULTS Activation was present in each section in all subjects. Subjective estimation of vapor intensity followed relative vapor pressure of stimuli presented (pyridine > amyl acetate > menthone) and were similar for both men and women. However, brain activation did not follow subjective responsiveness order but rather pyridine > menthone > amyl acetate, a pattern demonstrated by both men and women. Brain activation in women was consistently lower than in men for all vapors in all brain sections, although regions of activation did not differ. Activation occurred in regions previously recognized as associated with olfactory stimulation, including orbitofrontal and entorhinal cortex; however, extensive regions within frontal cortex including cingulate gyrus were also activated. Brain regions activated to odors considered pleasant or unpleasant did not differ. CONCLUSION The techniques used in this study demonstrated that brain activation to olfactory stimuli could be measured quantitatively such that differences between groups of subjects (in this case men and women) could be compared. Although localization of brain activation was not the major thrust of this study, activation to olfactory stimuli was found not only in brain regions previously associated with processing of olfactory information but also in several other areas of frontal cortex, in cingulate gyrus, and in several components of the limbic system. This is the first study in which activation in human brain parenchyma of normal humans to olfactory stimuli has been quantitated by fMRI.

Odor memory induces brain activation as measured by functional MRI.

PURPOSE Our goal was to use functional MRI (fMRI) to measure brain activation in response to imagination of odors in humans. METHOD fMR brain scans were obtained in 21 normal subjects (9 men, 12 women) using multislice FLASH MRI in response to imagination of odors of banana and peppermint and to the actual smells of the corresponding odors of amyl acetate and menthone, respectively, in three coronal sections selected from anterior to posterior temporal brain regions. Similar studies were obtained in two patients with hyposmia using FLASH MRI and in one patient with hyposmia using echo planar imaging, both before and after theophylline treatment, which returned smell function to or toward normal in each patient. Activation images were derived using correlation analysis, and ratios of areas of brain activated to total brain areas were calculated. RESULTS Activation was present in each section in all normal subjects and in each patient after imagination of each vapor. In normal subjects, brain activation in response to imagination of odors was significantly less than that in response to the actual smell of these odors, and activation following imagination of banana odor was significantly greater in men than in women, as was previously reported for the actual smell of the odor of amyl acetate. However, in relative terms, albeit at an absolute lower brain activation level, the ratio of brain activation by imagination of banana to activation by actual amyl acetate odor was about twice as high in women as in men. Before treatment, in patients with hyposmia, brain activation in response to odor imagination was greater than after presentation of the actual odor itself. After treatment, in patients with hyposmia in whom smell acuity returned to or toward normal, brain activation in response to odor imagination was not significantly different quantitatively from that before treatment; however, brain activation in response to the actual odor was significantly greater than that in response to imagination of the corresponding odor. Brain regions activated by both odor imagination and actual corresponding odor were similar and consistent with regions previously described as responding to odors. CONCLUSION These studies indicate that (a) odors can be imagined and similar brain regions are activated by both imagined and corresponding actual odors; (b) imagination of odors elicits quantitatively less brain activation than do actual smells of corresponding odors in normal subjects; (c) absolute brain activation in men by odor imagination is greater than in women for some odors, but on a relative basis, the ratio for odor imagination to actual smell in women is twice that in men; (d) odor imagination, once the odor has been experienced, is present, recallable, and capable of inducing a relatively constant degree of brain activation even in the absence of the ability to recognize an actual corresponding odor.

Taste memory induces brain activation as revealed by functional MRI.

PURPOSE Our goal was to use functional MRI (fMRI) to measure brain activation in response to imagination of tastes in humans. METHOD fMR brain scans were obtained in 31 subjects (12 men, 19 women) using multislice FLASH MRI and echo planar imaging (EPI) in response to imagination of tastes of salt and sweet in coronal sections selected from anterior to posterior temporal brain regions. Activation images were derived using correlation analysis, and ratios of areas of brain activated to total brain areas were calculated. Total activated pixel counts were used to quantitate regional brain activation. RESULTS Activation was present in each section in all subjects after imagination of each tastant. Activation was similar in response to imagination of either salt or sweet and was quantitatively similar to that previously reported in response to imagination of odors of banana and peppermint. Activation was similar in both men and women as opposed to previous results of odor memory in which activation in men was greater than in women. However, subjective responses of intensity of imagined tastes were significantly greater than those previously obtained for odor memory and were consistently, albeit not significantly, greater in women than in men, similar to results previously reported for odor memory. Brain regions activated in response to taste imagination were consistent with regions previously described as involved with actual taste perception in both humans and animals. Regional brain localization for salt and sweet memories could not be differentiated. CONCLUSION These studies indicate that (a) tastes can be imagined, (b) brain regions activated for taste imagination are consistent with regions previously described for actual taste perception, and (c) similar to odor memory for banana and menthone, regional brain localization for salt or sweet taste memories could not be differentiated.

Increased brain activation in response to odors in patients with hyposmia after theophylline treatment demonstrated by fMRI

PURPOSE Our goal was to demonstrate that medical therapy in patients with smell loss (hyposmia) that restored olfactory function toward or to normal could be verified and quantitated by functional MRI (fMRI) of brain and that visual representation of these changes could be used to identify these patients. METHOD Multislice FLASH MR or echo planar MR brain scans were obtained in four patients with hyposmia in response to three olfactory stimuli both before and after treatment with theophylline. Activation images were derived using correlation analysis, and ratios of brain area activated to total brain area were obtained. RESULTS Prior to treatment, all patients stated that they could not smell; these losses were confirmed by standard psychophysical tests. At this time, fMRI brain activation in response to odors was significantly less than that measured in normal volunteers and similar to activation measured previously in other patients with a similar type of hyposmia. After treatment, subjective smell function improved in three patients and no improvement occurred in one; results were confirmed by psychophysical tests. In each patient in whom smell acuity improved, brain activation in response to each odor increased in each section and mean activation increased significantly for each odor. Activation increased in all regions previously associated with olfactory stimulation and was particularly apparent in orbitofrontal cortex, frontal lobe component of cingulate gyri, temporal lobe gyri, and hippocampus. There also was consistent activation in superior, middle, and inferior frontal lobe gyri. There were no changes in brain activation after treatment in the patient in whom smell did not improve. CONCLUSION These results demonstrate that theophylline is an effective therapeutic agent to correct hyposmia in some patients with smell loss. These changes have been documented by fMRI brain scans using olfactory stimuli. This is the first study in which this type of objective improvement following medical treatment has been demonstrated in patients with hyposmia.

Lateralization of brain activation to imagination and smell of odors using functional magnetic resonance imaging (fMRI): left hemispheric localization of pleasant and right hemispheric localization of unpleasant odors.

Purpose Our goal was to use functional MRI (fMRI) of brain to reveal activation in each cerebral hemisphere in response to imagination and smell of odors. Method FMRI brain scans were obtained in 24 normal subjects using multislice fast low angle shot (FLASH) MRI in response to imagination of banana and peppermint odors and in response to smell of corresponding odors of amyl acetate and menthone, respectively, and of pyridine. Three coronal sections selected from anterior to posterior brain regions were used. Similar studies were obtained in two patients with hyposmia using FLASH MRI and in one patient with hyposmia using echo planar imaging (EPI) both before and after theophylline treatment that returned smell function to or toward normal in each patient and in two patients with birhinal phantosmia (persistent foul odor) and global phantogeusia (persistent foul taste) with FLASH and EPI fMRI before and after treatment with neuroleptic drugs that inhibited their phantosmia and phantogeusia. Activation images were derived using correlation analysis. Ratios of hemispheric areas of brain activation to total hemispheric brain areas were calculated for FLASH fMRI, and numerical counts of pixel clusters in each hemisphere were made for EPI studies. Total pixel cluster counts in localized regions of each hemispheric section were also obtained. Results In normal subjects, activation generally occurred in left (L) > right (R) brain hemisphere in response to banana and peppermint odor imagination and to smell of corresponding odors of amyl acetate and menthone. Whereas there were no overall hemispheric differences for pyridine odor, activation in men was R > L hemisphere. Although absolute activation in both L and R hemispheres in response to banana odor imagination and amyl acetate smell was men > women, the ratio of L to R activation was women > men. In hyposmic patients studied by FLASH fMRI, activation to banana odor imagination and amyl acetate smell was L > R hemisphere both before and after theophylline treatment. In the hyposmic patient studied with EPI before theophylline treatment, activation to banana and peppermint odor imagination and to amyl acetate, menthone, and pyridine smell was R > L hemisphere; after theophylline treatment restored normal smell function, activation shifted completely with banana and peppermint odor imagination and amyl acetate and menthone smell to L > R hemisphere, consistent with responses in normal subjects. However, this shift also occurred for pyridine smell, which is opposite to responses in normal control subjects. In patients with phantosmia and phantogeusia, activation to phantosmia and phantogeusia before treatment was R > L hemisphere; after treatment inhibited phantosmia and phantogeusia, activation shifted with a slight L > R hemispheric lateralization. Localization of all lateralized responses indicated that anterior frontal and temporal cortices were brain regions most involved with imagination and smell of odors and with phantosmia and phantogeusia presence. Conclusion Imagination and smell of odors perceived as pleasant generally activated the dominant or L > R brain hemisphere. Smell of odors perceived as unpleasant and unpleasant phantosmia and phantogeusia generally activated the contralateral or R > L brain hemisphere. With remission of phantosmia and phantogeusia, hemispheric activation was not only inhibited, but also there was a slight shift to L > R hemispheric predominance. Predominant L > R hemispheric differences in brain activation in normal subjects occurred in the order amyl acetate > menthone > pyridine, consistent with the hypothesis that pleasant odors are more appreciated in L hemisphere and unpleasant odors more in R hemisphere. Anterior frontal and temporal cortex regions previously found activated by imagination and smell of odors and phantosmia and phantogeusia perception accounted for most hemispheric differences.

Taste and smell phantoms revealed by brain functional MRI (fMRI).

PURPOSE Our goal was to demonstrate the appearance of phantom tastes and smells (phantageusia and phantosmia, respectively) by use of functional MRI (fMRI) of the brain and to demonstrate the efficacy of drug treatment that inhibited both the subjective presence of these phantoms and the fMRI brain activation initiated by these phantoms. METHOD Multislice FLASH MR or echo planar MR brain scans were obtained in two patients with phantageusia and phantosmia in response to memory of two tastants (salt and sweet); memory of two odors (banana and peppermint); actual smell of amyl acetate, menthone, and pyridine; and memory of phantom tastes and smells before and after treatment with thioridazine and haloperidol. Activation images were derived using correlation analysis, and ratios of brain area activated to total brain area were obtained. RESULTS Prior to treatment, both patients experienced persistent birhinal and global oral obnoxious tastes and smells in the absence of any external stimulus. The fMRI response to memory of phantoms was activation in sensory-specific brain regions for taste and smell, respectively. fMRI activation was greater than for memory of any tastant or odorant or for actual smell of any odor. After treatment with thioridazine or haloperidol, which successfully inhibited each phantom in each patient, fMRI response to phantom memory was significantly inhibited and was significantly lower than for memory of any tastant or odorant or actual smell of any odorant. CONCLUSION These results demonstrate that (a) phantom taste and smell can be revealed by fMRI brain activation, (b) brain activation in response to taste and smell phantoms is localized in sensory-specific brain regions for taste and smell, respectively, (c) brain activation in response to memory of each phantom initiated the greatest degree of activation we had previously measured, and (d) treatment with thioridazine or haloperidol inhibited both the presence of each phantom and its associated fMRI brain activation. This is the first study in which phantom tastes and smells have been demonstrated by an objective technique and treatment that inhibited the phantoms was characterized by objective inhibition of fMRI activation. These two patients represent a relatively common group that may be classified as having primary phantageusia and phantosmia distinct from those with phantoms or auras secondary to neurological, migrainous, psychiatric, or other causes.

Functional MRI of congenital hyposmia: brain activation to odors and imagination of odors and tastes.

Purpose Our goal was to use functional MRI (fMRI) to define brain activation in response to odors and imagination (“memory”) of odors and tastes in patients who never recognized odors (congenital hyposmia). Method Functional MR brain scans were obtained in nine patients with congenital hyposmia using multislice echo planar imaging (EPI) in response to odors of amyl acetate, menthone, and pyridine and to imagination (“memory”) of banana and peppermint odors and to salt and sweet tastes. Functional MR brain scans were compared with those in normal subjects and patients with acquired hyposmia. Activation images were derived using correlation analysis, and ratios of areas of brain activated to total and hemispheric brain areas were calculated. Total and hemispheric activated pixel counts were used to quantitate regional brain activation. Results Brain activation in response to odors was present in patients with congenital hyposmia. Activation was significantly lower than in normal subjects and patients with acquired hyposmia and did not demonstrate differential vapor pressure-dependent detection responsiveness or odor response lateralization. Regional activation localization was in anterior frontal and temporal cortex similar to that in normal subjects and patients with acquired hyposmia. Activation in response to presented odors was diverse, with a larger group exhibiting little or no activation with localization only in anterior frontal and temporal cortex and a smaller group exhibiting greater activation with localization extending to more complex olfactory integration sites. “Memory” of odors and tastes elicited activation in the same central nervous system (CNS) regions in which activation in response to presented odors occurred, but responses were significantly lower than in normal subjects and patients with acquired hyposmia and did not lateralize. Conclusion Odors induced CNS activation in patients with congenital hyposmia, which distinguishes olfaction from vision and audition since neither light nor acoustic stimuli induce CNS activation. Odor activation localized to anterior frontal and temporal cortex, consistent with the hypothesis that olfactory pathways are hard-wired into the CNS and that further pathways are undeveloped with primary olfactory system CNS connections but lack of secondary connections. However, some patients exhibited greater odor activation with response localization extending to cingulate and opercular cortex, indicating some olfactory signals impinge on and maintain secondary connections consistent with similar functions in vision and audition. Activation localization of taste “memory” to anterior frontal and temporal cortex is consistent with CNS plasticity and cross-modal CNS reorganization as described for vision and audition. Thus, there are differences and similarities between olfaction, vision, and audition, the differences dependent on unique qualities of olfaction, perhaps due to its diffuse, primitive, fundamental role in survival. Response heterogeneity to odors may reflect heterogeneous genetic abnormalities, independent of anatomic or hormonal changes but dependent on molecular abnormalities in growth factor function interfering with growth factor/stem cell interactions. Patients with congenital hyposmia offer an unique model system not previously explored in which congenital smell lack as measured by fMRI is reflective of congenital dysfunction of a major sensory system.

Mapping Brain activation to odorants in patients with smell loss by Functional MRI

PURPOSE Our goal was to use functional MRI (fMRI) to develop an objective, noninvasive technique by which patients with smell loss can be identified, their abnormalities quantitated, their results compared with findings in normal subjects, and visual representation of their CNS pathology obtained. METHOD Functional MR brain scans were obtained in eight patients with hyposmia in response to three olfactory stimuli (pyridine, menthone, amyl acetate) in three coronal brain sections selected from anterior to posterior temporal brain regions using multislice FLASH MRI. Results were compared with similar studies performed in 17 normal subjects. Activation images were derived using correlation analysis, and ratios of area of brain activated to total brain area were obtained. RESULTS Brain activation to each stimulus was lower in each section in patients compared with normal subjects and reached statistical significance for mean activation for each odor and in six of the nine individual sections studied. Activation in patients was found in regions previously associated with CNS processing of olfactory stimuli in normal subjects, but activation in patients was much less, particularly in inferior frontal and cingulate gyral regions of frontal cortex and in regions of medial and posterior temporal cortex. CONCLUSION These results demonstrate quantitative CNS changes in smell function in response to olfactory stimuli in patients with hyposmia, demonstrate a novel, objective method by which these patients can be identified, and provide maps of the CNS changes associated with their smell loss.

Brain gamma-aminobutyric acid levels are decreased in patients with phantageusia and phantosmia demonstrated by magnetic resonance spectroscopy

Background: Olfactory and gustatory hallucinations (phantosmias and phantageusias, respectively) are sensory distortions that commonly follow losses of olfactory and gustatory acuity (hyposmia and hypogeusia, respectively). The biochemical basis of these hallucinations is unclear. Functional magnetic resonance imaging has been used previously to demonstrate widespread and robust central nervous system (CNS) activation to memories of these sensory distortions in patients with these symptoms. In this study, possible CNS mechanisms responsible for these distortions were evaluated using magnetic resonance spectroscopy, because this technique has been used to measure various CNS metabolites in patients with neurologic disorders. Methods: Forty-seven subjects were studied: 28 normal volunteers (13 men and 15 women) and 19 patients (8 men and 11 women) with persistent oral global phantageusia and/or birhinal phantosmia studied before any treatment. Four patients (1 man and 3 women) were studied before and after pharmacologic treatment that reduced the severity of their sensory distortions. All subjects were studied in a Signa 1.5-T magnetic resonance scanner with a quadrature head coil using a modified standard 2-dimensional J-point resolved excitation in the steady state (PRESS) sequence by which gamma-aminobutyric acid (GABA), glutamic acid, choline, N-acetylaspartate, and creatine (Cre) were measured in various CNS regions. Results were expressed using Cre as a denominator to determine ratios for each measurement. Differences were defined between normal subjects and patients before treatment and in patients before and after successful pharmacologic treatment. Results: Before treatment, GABA levels in several CNS regions were lower in patients than in normal volunteers and were the only biochemical changes found; significantly lowered GABA levels were found in the cingulate, right and left insula, and left amygdala. No differences between patients and normal volunteers were found in any of the metabolites in the posterior occipital region. After treatment that inhibited sensory distortions, CNS GABA levels increased in the cingulate, insula, and amygdala but significantly only in the left insula and in the right and left amygdala. After this successful treatment, no change in any biochemical parameter was found in the posterior occipital region. Conclusions: These results indicate that decreased brain GABA levels can serve as biochemical markers of phantageusia and/or phantosmia in patients with these distortions and are the first biochemical changes in the CNS that reflect these sensory changes. After successful treatment of these distortions, CNS GABA levels increased to levels at or near normal, consistent with functional remission of these symptoms. These results substantiate a role for CNS GABA in the generation and inhibition of these sensory hallucinations. Although the underlying biochemical mechanism(s) for the generation of these decreased GABA levels are complex, because similar types of sensory hallucinations occur as auras or prodromata of epileptic seizure and migraine activity, these results suggest that there may be common biochemical changes among these disorders.

Rapid imaging of olfaction by functional MRI (fMRI): identification of presence and type of hyposmia.

PURPOSE Our goal was to develop a rapid, simple, near-real-time method of functional MRI (fMRI) to measure brain activation in response to olfactory stimuli, to use it to identify patients with smell loss (hyposmia), and to differentiate their types of hyposmia. METHOD fMRI was obtained in 16 patients with Type I hyposmia (who could detect but not recognize odors), 5 patients with Type II hyposmia (who could both detect and recognize odors, albeit with less than normal acuity), and 2 volunteers with normal olfactory acuity by use of a rapid echo planar imaging technique in which one coronal brain section from the anterior cortical region was studied and a single olfactory stimulus was used. Actual scanning time performed by a variation of methods previously published required 26 s. Three patients with Type I hyposmia were treated with theophylline 250-500 mg for 4-6 months and were studied before and after treatment. RESULTS Brain activation in response to olfactory stimuli was demonstrated using a new, rapid, and simple fMRI technique. Patients with Type I hyposmia had less activation than patients with Type II hyposmia. Both patient groups had less activation than normal volunteers. Activation in patients with Type I hyposmia was essentially absent from regions of the middle frontal, orbitofrontal, and temporal cortex and was totally absent in regions of inferior frontal, insular, and cingulate cortex. Activation in patients with Type II hyposmia was greatest in the middle frontal cortex and the orbitofrontal cortex bilaterally and was present in regions of inferior frontal, temporal, and cingulate cortex. Each patient with Type I hyposmia treated with theophylline had improved smell function to Type II hyposmia and after treatment demonstrated activation in inferior frontal and cingulate cortex bilaterally, whereas before treatment, no activation in these regions was apparent. CONCLUSION We describe a simple, rapid technique that can be used in a practical clinical setting to identify patients with hyposmia and to differentiate patients with different types of olfactory loss. These studies confirm the presence and classification of patients with Type I and Type II hyposmia. Results of this study suggest that regions of the frontal cortex may act to guide or direct olfactory signals to other brain areas such as temporal and cingulate regions. Although these latter regions are involved with olfactory recognition, their role in olfactory memory, olfactory meaning, and attention needs to be considered.