Imaging of the Olfactory System David M. Yousem, Kader Karli Oguz, and Cheng Li The olfactory system consists of the primary olfactory nerves in the nasal cavity, the olfactory bulbs and tracts, and numerous intracranial connections and pathways. Diseases affecting the sense of smell can be located both extracranially and intracranially. Many sinonasal inflammatory and neoplastic processes may affect olfaction. Intracranially congenital, traumatic, and neurodegenerative disorders are usually to blame for olfactory dysfunc- tion. The breadth of diseases that affect the sense of smell is astounding, yet the imaging ramifications have barely been explored. Copyright © 2001 by W,B. Saunders Company HE ANATOMY AND PHYSIOLOGY of ol- olfactory tract to enter the brain just lateral and T faction have rarely been the subject of arti- anterior to the optic chiasm. Fibers continue into cles in the radiologic literature, in part because this medial and lateral olfactory striae to septal nuclei special sense is the least appreciated one in the at the base of the brain just inferior and anterior to neurosciences. All too often the olfactory system the rostrum of the corpus callosum. Axons from is ignored in the usual clinical evaluation of a mitral and tufted cells project to central brain patient, hence, the common recording that "cranial limbic system components including the pyriform nerves II to XII are intact." Nonetheless, to the and entorhinal cortex and adjacent corticomedial astute clinician and conscientious neuroradiologist, amygdala (which together form the uncus), the knowledge of the ramifications of olfactory dys- ventral striatum, the parahippocampus area, and function is very rewarding. Therefore, this article the anterior olfactory nuclei. From these areas initially focuses on the functional anatomy of the there are widespread interconnections with many sense of smell and then addresses those protean parts of the brain, including the mediodorsal thal- entities that affect cranial nerve I (see Table 1). amus, hypothalamus, orbitofrontal and dorsolateral Both peripheral and central diseases that impact on frontal cortex, temporal cortex, and other areas of the sense of smell are explored. the limbic system (Fig 2). The sizes of the olfactory bulbs and tracts ANATOMIC IMAGING (NORMAL) (OBTs) have recently been reported in the litera- Smells are perceived in the upper nasal cavity by ture. 1 The left and right OBT volumes peak in the olfactory neuroepithelial receptors. The primary fourth decade of life, but show considerable vari- olfactory nerves pierce the cribriform plate to stim- ation within each subject group (see Table 2). ulate and synapse with olfactory bulb nuclei. In The mean left and fight OBT volumes and the truth, the true olfactory nerves are located in the mean combined volume show a decline in the nasal cavity and these should be termed cranial seventh and eighth decade that parallels a decline nerves I (CN I). Instead, the olfactory bulb and in scores by the patients on the University of tract, which is a secondary neuron, is usually Pennsylvania Smell Identification Test (UPSIT) referred to as the first cranial nerve (Fig 1). scores. Thus, as the volume declines, function From the olfactory bulb, fibers travel in the declines. A Kruskal-Wallis test, performed to de- tect differences between the means of OBT vol- umes between decades showed borderline signifi- From the Russell H. Morgan Department of Radiology and cance at P = .05 for the left OBT, but not the right. Radiological Sciences, Johns Hopkins Hospital, Baltimore, The biggest differences noted were between the MD; and the Smell and Taste Center, Department of Otorhi- fourth and seventh decade for fight and left OBTs. nolaryngology, Head and Neck Surgery, Hospital of the Uni- By using regression analysis based on the gener- versity of Pennsylvania, Philadelphia, PA. Address reprint requests to David M. Yousem, MD, Director alized estimating equations model, a OBT volumes of Neuroradiology, The Russell H. Morgan Department of show significant decreases between decades of life Radiology and Radiological Sciences, Johns Hopkins Hospital, (P = .027). 600 North Wolfe Street, Phipps B-112, Baltimore, MD 21287; As measured in large population studies by us- e-mail: [email protected] ing the UPSIT, there is an initial increase in odor Copyright © 2001 by W.B. Saunders Company 0887-2171/01/2206-0001535.00/0 identification capability to the third decade of life, doi: l O.1053/sult.2001.28800 at which time the UPSIT scores plateau? At about 456 Seminars in Ultrasound, CT, and MRI, Vol 22, No 6 (December), 2001: pp 456-472 IMAGING OF THE OLFACTORY SYSTEM 457 Table 1, Causes of Olfactory Dysfunction Primary Olfactory Apparatus Conductive Disorders Lesions NeurodegenerativeDisorders Miscellaneous Sinusitis Sinusitis Alzheimer's disease Vasculitis Polyposis Traumatic injury Parkinson's disease infarction Sinonasal cavity neoplasms Congenital aplasia, hypoplasia Huntington's disease Surgery Olfactory neuroblastoma Schizophrenia Cocaine Multiple sclerosis Toxins Temporal lobe epilepsy Viral infections Korsakoff's disease Amyotropic lateral sclerosis age 60, however, a decline in median UPSIT the brain associated with olfaction. Koizuka et a116 scores begins to be seen across patients to the point have noted bilateral increased cerebral blood flow that nearly 75% of individuals over 80 years old in the inferior frontal lobes, piriform cortex, and score 19 or less on the 40-item UPSIT test. 4 These orbitofrontal regions after phenyl ethyl alcohol patients are severely hyposmic or anosmic. Similar olfactory stimulation. Wexler et al t7 showed sim- changes in odor threshold values for phenyl ethyl ilar areas of activation in these and other limbic alcohol are noted, however, the drop-off starts to structures when odors were presented to normal be seen at about age 40 in men and 60 in women. 5 volunteers. Smokers and men tend to have higher thresholds Yousem et a118 have performed olfactory-stim- than nonsmokers and women, respectively. A ulated functional magnetic resonance imaging change in the perception of pleasantness of taste (OSfMRI) in 18 normal subjects between 29 and also occurs with age. 6'7 This effect can be elimi- 43 years of age by using a homemade olfactometer nated when the nostrils are occluded and the olfac- in which odors from 3 sources were alternated at tory influence on taste is nullified.* 5-second intervals for 30 seconds during the on Several theories have been espoused as to why stimulus and room air was used for 30 seconds for the sense of smell declines with age and include: the off stimulus. They found that odorants that (1) reduction in volume of the sensory neuroepi- stimulated only CN I had extensive activation that thelium caused by cumulative effects of viral in- was localized to the orbitofrontal region (Brod- fections, (2) replacement of the olfactory neuro- mann area 11) and the cerebellum. The volume of epithelium with nonsensory columnar epithelium activation was greater in the right orbitofrontal in the olfactory clefts of the nose, (3) diminution in region than the left (Table 3, Fig 3). central neurotransmitters (eg, norepinephrine), (4) Some odorants not only stimulate the olfactory reduction in patency of the airway, (5) variations in system but also activate trigeminal neurons that the nasal cycle with age, or (6) reduction in resis- abound in the nasal cavity. Stimulants such as tance of the olfactory neuroepithelium to infectious carbon dioxide that "sting or burn the nose" tend to or toxic insults. 4"5'9-a5 have more fifth-cranial nerve input. When Yousem The finding of a decline in the human OBT et a118 applied an fMRI paradigm with odorants volume with advancing age lends credence to the that also induce trigelmnal nerve stimulation they idea of a reduced amount of afferent input to the found wider activation of many different areas, olfactory bulbs and tracts from the olfactory neu- including visual and cingulate areas (Table 4, Fig roepithelium and ciliated nerves that pierce the 4). Yousem et al ~8 found that there was accommo- cribriform plate. A decrease in central neurotrans- dation (diminution of brain activation) to pleasant mitters is a less plausible explanation for the quan- olfactory nerve-mediated odors, but amplification titative findings because the predominant effect (an increase in brain activation) to repeated testing seen is at the bulb and tract level, not at the cortical with trigeminally mediated odors. level. Fulbright et a119 reported that pleasant odors lateralized to the left insula region whereas the FUNCTIONAL IMAGING (NORMAL) unpleasant odors were more left frontal. Functional magnetic resonance imaging (fMRI) In a study of the effects of age on fMRI, 2° 5 has been used in an attempt to localize regions of older right-handed patients (mean age 73.2 years) 458 YOUSEM, OGUZ, AND LI Fig 1. Normal olfactory bulb and tract. (A) The olfactory bulbs (open arrows), olfactory sulci (arrowheads), and gyri rectus (g) are well seen on coronal Tl-weighted surface coil images (e, ethmoid sinuses). (B) At a more posterior cut, the open arrows now show the olfactory tracts sitting in the olfactory sulci (arrowheads). (C) Even more posteriorly, normal small tracts (arrows) are seen in the sulci. and 5 younger right-handed patients (mean age of ized to a standard atlas, and individual and group 23.8 years) underwent blood oxygenation level statistical parametric maps (SPMs) were generated dependent (BOLD) fMRI scans that used binasal for each task. The SPMs (P < .01) of volumes of olfactory nerve stimulation. The data were normal- activation and distribution of cluster maxima were IMAGING OF THE OLFACTORY SYSTEM 459 A B C D Olfactory Olfactory tract bulb Olfactory Fig 2. Connections to cortex from the olfactory system (diagram or graph). Organization of the human olfactory system. (A) Peripheral and central components of the olfactory pathway. (B) Enlargement of region boxed in (A) showing the relationship between the olfactory epithelium, containing the olfactory receptor neurons, and the olfactory bulb (the central target of olfactory receptor neurons).