Introduction

The etiologies of olfactory dysfunction include congenital causes, ageing, idiopathic changes, infections of the upper respiratory tract (URTI), sinonasal disease (SND), traumatic brain injury (TBI), and neurologic illnesses including Alzheimer’s disease, multiple sclerosis and Parkinson’s disease (Wiesmann et al. 2001; Barresi et al. 2012). Olfactory dysfunction can affect all areas of life—from preparation and enjoyment of foods to the appreciation of the fragrances of flowers, detection and avoidance of hazardous odours, maintenance of personal hygiene, and the social or intimate interactions with others (Stevenson 2009; Croy et al. 2014).

The sense of smell allows us to identify the chemical nature in the surroundings. Sensory neurons in the nose detect odour molecules and transmit signals to the olfactory bulb (OB), a structure in the forebrain where initial odour processing occurs (Han et al. 2019). The OB collects the sensory afferents of the olfactory receptor cells located in the olfactory neuroepithelium. The axonal projections are then conveyed with a relay in the OB to the piriform cortex through the olfactory tract (Gottfried 2010). The piriform cortex projects to multiple brain regions within the limbic system. This system is directly connected with the frontal cortex through pathways to the posterior orbitofrontal cortex (OFC) as well as indirectly via the mediodorsal thalamic nucleus (Carmichael et al. 1994; Illig 2005). The OB is closely related to the olfactory sulcus (OS), located in the frontal lobe (Rombaux et al. 2009b). Previous studies suggest the OB volume (Mazal et al. 2016; Shehata et al. 2018; Yousem et al. 1996b), and OS depth (Miao et al. 2015) change according to different types of olfactory disorders suggesting that peripheral olfactory input influences the OB volume, in which OB volume is smaller, and OS depth is shallow in anosmic patients. Little is known about cortical brain areas beyond the OB and OS. Only a few studies reported pattern and variability of grey matter (GM) and white matter (WM) in patients with olfactory loss (Bitter et al. 2010a, b; Peng et al. b, 2021; Yap et al. 2021). The search was performed to identify studies reporting congenital anosmia and acquired anosmia using magnetic resonance imaging (MRI). Article search was conducted between the earliest record and 26 July 2021. Search terms were as follow: ((((((((((anosmia)) OR (congenital anosmia)) OR (acquired anosmia)) OR (smell impairment)) OR (olfactory loss)) OR (olfactory dysfunction)) OR (olfactory deficit)) OR (smell blindness)) OR (smell dysfunction)) AND ((( (Magnetic resonance imaging)) OR (MRI)) OR (MR Imaging)). We also manually checked for related articles in references and citations through the Google Scholar database. There was no limitation on the publication date. All records were grouped into a final database after removing duplicates, followed by screening the titles and abstracts and finally full-text article screening and eligibility by HAM and NY, independently, the details of the selected studies are tabulated in Fig. 1. Consensus for eligibility was reached through discussion. We used an assessment tool from the National Heart, Lung and Blood Institute, Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies, to assess the quality of included studies https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools.

Fig. 1
figure 1

Diagram of the search process for studies included in the present systematic review

Full size image

Inclusion criteria and exclusion criteria

Original studies reported in peer-reviewed journals describing research on structural brain changes in anosmia using MRI were included. We included studies describing research on patients with anosmia (congenital and acquired) that include studies that used standardised measures of olfactory function, e.g., the measurement of odour threshold, discrimination and identification (TDI score) (Hummel et al. 2017), and assessing any olfactory component; OB volume, OS depth, WM, and GM. We excluded all review articles as well as case reports and case series studies. We also exclude functional MRI (fMRI), diffusor tensor imaging (DTI), electro-encephalography (EEG), and magneto-encephalography (MEG) studies. We also exclude olfactory dysfunction in relation to neurodegenerative diseases and neuropsychiatric disorders such as; Alzheimer's and syndromes such as Kallmann and Bardet Biedl syndromes. Finally, we exclude anosmia in patients with SARS-CoV-2 or COVID-19 infection because findings on COVID-19 are currently develo** and unclear. We estimate that olfactory loss due to COVID-19 requires an independent review due to its emerging nature. Following the removal of duplicates and citations from non-English language journals, those evidently outside the review’s scope were rejected. From the eligible studies, the following variables were recorded: year of publication, country, study author(s), analysis mode, participants’ demographics, including age, handedness, duration of olfactory loss, psychophysical and physiological tests, and principal findings.

Results

Study demographics and details

Table 1 provides a summary of demographic information of anosmic patients with various etiologies, including congenital anosmia, idiopathic olfactory loss (IOL), infections of the upper respiratory tract (URTI), and traumatic brain injury (TBI). Across 28 studies, 2024 participants have investigated; 1639 patients and 385 healthy controls. Generally, studies have reasonable quality, as shown in Supplementary A. Sample size calculations were rarely mentioned, and the number of anosmic patients reported was between 3 and 378 patients per study. None of the studies was blinded due to the nature of the studies, which require the direct involvement of personnel-in-charge. Gender of participants were reported in some of the studies and indicated slightly more females (anosmia: males = 581, females = 625 and not mentioned = 433, healthy controls: male = 173, female = 196 and not mentioned = 16). The studies were conducted in multiple countries, including South Korea, The Netherlands, Germany, Belgium, PR China, the USA, and Taiwan. The age of the participants ranged from 9 to 73 years old. Seventeen studies compared anosmic patients to age-and sex-matched healthy controls. However, no study conducted separate analyses based on age and gender. Seven studies exclusively accrued right-handed participants (Bitter et al. 2010a, b; Peng et al. 2011).

Limitations in the literature

The heterogeneity of the findings was high due to the differences in many factors. The results show morphological variations, particularly in GM and WM atrophy. For example, both congenital and acquired anosmia studies reported inconsistency in brain areas and laterality. Even though some of the studies reported similar aetiology, their findings were at times contradictory. Six studies reported congenital anosmia; however, only three studies reported larger volumes of GM and WM compared to healthy controls in a few areas related to olfactory processing. Most importantly, the areas involved and the laterality of the involved areas were varied from one study to another. Furthermore, most of the existing studies also have a very small sample size, and future research requires further investigation in a larger sample.

Future directions for research in this area

We would like to suggest more comprehensive research in the future for both congenital and acquired olfactory loss. This includes more regions being considered, more comprehensive measures of anosmia, and the usage of MRI methods like DTI. In the current study, we proposed that the neural mechanism of brain alterations is different between aetiology, but the underlying details of these differences are unclear. More importantly, congenital anosmia shows larger GM and WM volume than HC, which is not found in acquired anosmia. This finding is fascinating, but the current data show discrepancies in terms of areas involved and brain laterality.

Conclusion

The present review suggests that MRI evaluations of OB volume could objectively diagnose olfactory dysfunction in patients with subjective olfactory loss. However, because the correlation between OS depth and olfactory dysfunction is not apparent and contradictory, a combined OB volume and OS depth evaluation is suggested. We also observed the right dominance of OB volume and OS depth, which is in line with the idea that the right hemisphere is relatively more important for olfactory processing than the left hemisphere. The present review also observed both primary and secondary olfactory areas show GM and WM alterations and conclude that brain alteration is more pronounced with longer disease duration. Finally, congenital anosmia shows larger GM and WM volumes in a few regions in primary and secondary olfactory areas. This is opposite to the presence of smaller GM and WM volumes in patients with acquired olfactory loss. The present study suggests that the difference in the volume and thickness of GM and WM between congenital and acquired anosmia is due to different mechanisms responsible for the olfactory dysfunction. We further suggest that the mechanisms underlying congenital anosmia differ from those involved in acquired loss. The mechanism behind these structural and neural changes are likely to be multifactorial. This result motivates further neuroimaging research into the pathophysiology of lifelong olfactory dysfunction.