Changes in Structural Covariance among Olfactory-related Brain Regions in Anosmia Patients

Abstract

Anosmia, characterized by the loss of smell, is associated not only with dysfunction in the peripheral olfactory system but also with changes in several brain regions involved in olfactory processing. Specifically, the orbitofrontal cortex is recognized for its pivotal role in integrating olfactory information, engaging in bidirectional communication with the primary olfactory regions, including the olfactory cortex, amygdala, and entorhinal cortex. However, little is known about alterations in structural connections among these brain regions in patients with anosmia. In this study, high-resolution T1-weighted images were obtained from participants. Utilizing the volumes of key brain regions implicated in olfactory function, we employed a structural covariance approach to investigate brain reorganization patterns in patients with anosmia (n=22) compared to healthy individuals (n=30). Our structural covariance analysis demonstrated diminished connectivity between the amygdala and entorhinal cortex, components of the primary olfactory network, in patients with anosmia compared to healthy individuals (z=-2.22, FDR-corrected p=0.039). Conversely, connectivity between the orbitofrontal cortex—a major region in the extended olfactory network—and amygdala was found to be enhanced in the anosmia group compared to healthy individuals (z=2.32, FDR-corrected p=0.039). However, the structural connections between the orbitofrontal cortex and entorhinal cortex did not differ significantly between the groups (z=0.04, FDR-corrected p=0.968). These findings suggest a potential structural reorganization, particularly of higher-order cortical regions, possibly as a compensatory effort to interpret the limited olfactory information available in individuals with olfactory loss.

INTRODUCTION

Anosmia, the loss of the sense of smell, can significantly impact the psychological and behavioral aspects of human life, given that smell is closely linked to emotion, memory, social interaction, and hedonic experience [1]. Beyond dysfunctions in the peripheral olfactory system, anosmia is associated with structural and functional changes in several brain regions involved in olfactory processing [2-4]. Consequently, anosmia can increase the risk of developing psychiatric conditions such as depression and anxiety [5, 6]. Additionally, anosmia is recognized as both an early symptom and a risk factor for neurological conditions like dementia and Alzheimer’s disease [7].

Previous neuroimaging studies, including our own [8], have demonstrated the structural brain alterations associated with anosmia [3, 9]. These studies have focused on regions within both the primary olfactory network, such as the olfactory cortex, amygdala, and entorhinal cortex, and the secondary olfactory network, including the orbitofrontal cortex [3, 9, 10]. Research has consistently identified volume loss in the olfactory bulb and olfactory cortex—critical for olfactory perception—as being directly linked to olfactory dysfunction [3, 4]. However, findings on changes in other higher-order brain structures implicated in the modulation of olfactory information, remain inconsistent [4]. These inconsistencies may reflect the complex and interactive brain responses to insufficient olfactory perception.

Specifically, brain regions including the amygdala, entorhinal cortex, and orbitofrontal cortex are recognized for their essential role in olfactory processing, particularly through their complex and reciprocal interactions [3, 10, 11]. Therefore, moving beyond the traditional focus on individual brain structures in anosmia research, assessing the connectivity and interactions among these key brain regions is essential for enhancing our understanding of the complex brain mechanisms underlying anosmia.

Structural covariance analysis has recently been utilized to investigate the structural connections and organization among various brain regions [12, 13]. This method involves assessing the covariation patterns across different brain regions by analyzing the correlation of regional volumes or thickness measures within a group. Such covariance patterns are indicative of structural connections between the corresponding brain regions and can be influenced by a range of factors, including developmental or genetic factors, as well as acquired factors like the impact of sensory deprivation or stimulation [14, 15]. Accordingly, structural covariance analysis has been successfully used to study the influence of various neuropsychiatric diseases on connections between brain regions involved in cognitive and emotional processing [16, 17]. However, to the best of our knowledge, this approach has not yet been applied to explore how anosmia affects structural connections within the brain regions critical for olfactory processing.

Therefore, we employed the structural covariance analysis to compare the structural covariance patterns of the key brain regions—specifically focusing on the amygdala, entorhinal cortex, and orbitofrontal cortex—that play a crucial role in olfactory processing, between patients with anosmia and healthy individuals. We hypothesize that anosmia will lead to altered structural covariance among these olfactory-related brain regions, potentially reflecting disruptions in olfactory function or compensatory mechanisms.

MATERIALS AND METHODS

Participants

Participants of this study were derived from a previously reported study [8], comprising 22 patients with anosmia and 30 healthy individuals. Anosmia is defined as a persistent inability to perceive odors for a duration of at least one month and was confirmed through nasal endoscopic examinations and olfactory function evaluations. Olfactory function was assessed using the butanol threshold test [18, 19]. This test involves presenting participants with a series of ten dilutions of 4% N-butanol in mineral oil. Starting with the most dilute concentration, participants were asked to distinguish between butanol and mineral oil. The test progressed to more concentrated solutions until the participant correctly identified butanol in four successive attempts, thereby establishing their olfactory sensitivity threshold. Anosmia was diagnosed in individuals with thresholds at levels 0~1 [19].

Individuals with major medical, neurological, or psychiatric disorders, or those with contraindications for magnetic resonance imaging (MRI) were excluded from the study.

The research protocol was approved by the Institutional Review Board of Ewha W. University, and all participants provided written informed consent prior to participation.

Image data acquisition and volume measurements

High-resolution T1-weighted MRI images were acquired using a 3.0 Tesla Magnetic Resonance scanner (Achieva, Philips Medical Systems, Best, the Netherlands). Due to an update in imaging parameters during the research, two scanning protocols for T1-weighted images were employed. Protocol 1 was administered to a portion of the anosmia group (n=16), while protocol 2 was applied to the rest of the anosmia participants (n=6) and all control subjects (n=30). Protocol 1 settings included: a repetition time (TR) of 7.4 ms, echo time (TE) of 3.4 ms, a flip angle of 8°, a slice thickness of 1.0 mm without any interslice gap, a field of view (FOV) of 220×220 mm², and a matrix size of 220×209. Protocol 2 adjusted the FOV and matrix size, setting them to 224 mm and 224×216, respectively. In addition, fluid-attenuated inversion recovery (FLAIR) imaging was performed to identify any structural abnormalities within the brain, utilizing parameters such as a TR of 4,800 ms, TE at the shortest, an inversion time of 1,650 ms, a FOV of 240×240 mm², and a matrix size of 216×216.

Volume measurements of the amygdala, entorhinal cortex, and orbitofrontal cortex—critical regions implicated in olfactory processing—were conducted using the FreeSurfer 7.2 (http://www.surfer.nmr.mgh.harvard.edu/). The data processing workflow included the following steps: motion correction, skull stripping, automated Talairach transformation, parcellation of white and grey matter, intensity normalization, tessellation of brain boundaries, topology correction, and surface deformation [20, 21]. An experienced researcher carefully reviewed each step of cortical reconstruction and the generated images.

Structural covariance analysis using volume data

We employed structural covariance analysis to estimate connectivity strength among the amygdala, entorhinal cortex, and orbitofrontal cortex, based on regional volume data [12, 22, 23] (Fig. 1). For this analysis, partial correlation analysis was performed to calculate correlation coefficients between each pair of brain regions, controlling for intracranial volume. These partial correlation coefficients reflect the group-level connectivity strength among the specified brain regions. We then converted these coefficients into z-values using Fisher’s r-to-z transformation [24]. To assess differences in the connectivity strength between the groups, z-test statistics were applied. This analysis specifically determined whether the group-level correlation coefficient in the anosmia group was significantly greater or smaller than those in the control group, taking into account the sample size of each group [24, 25]. We applied a false discovery rate (FDR) correction set at q=0.05 to adjust for multiple comparisons [26]. The statistical analysis was conducted using R software, version 4.3.2 (the R Foundation, Vienna, Austria).

Fig. 1.

Key brain regions implicated in olfactory processing: a focus of our study.

Statistical analysis

Demographic variables were compared between groups using independent t-test for continuous variable and chi-square tests for categorical variables. Statistical significance was set at a two-tailed p<0.05. All statistical analyses were performed using Stata SE version 16.0 (StataCorp LP, College Station, Texas, USA).

RESULTS

Demographic characteristics, including age, gender, and handedness, did not show significant differences between the groups. The median duration of anosmia was 4 years, ranging from 1 to 21 years. Chronic rhinosinusitis was the primary cause of anosmia (n=15, 68.2%), while idiopathic causes accounted for 31.8% of cases (n=7). Participant characteristics are provided in Table 1.

Table 1.

Characteristics of study participants

Characteristics	Control (n=30)	Anosmia (n=22)	p
Age, years, mean (SD)	48.5 (10.3)	48.5 (10.3)	0.25
Female, n (%)	9 (30.0)	9 (40.9)	0.41
Right handedness, n (%)	28 (93.3)	20 (90.9)	0.75
Duration of olfactory loss, years, mean (SD)	NA	5.7 (5.1)	NA

Structural covariance analysis demonstrated the group-level connectivity strength among the amygdala, entorhinal cortex, and orbitofrontal cortex for both the control (Fig. 2A) and anosmia (Fig. 2B) groups (Table 2). We compared the connectivity strength between the groups using z-test statistics, as detailed in Table 3 and Fig. 2C.

Fig. 2.

Structural covariance analysis of key brain regions implicated in olfactory processing. Group-level correlation matrices for the control (A) and anosmia (B) groups illustrate the connectivity strength among the amygdala, entorhinal cortex, and orbitofrontal cortex. The z-value matrix (C) represents comparisons of correlation coefficients between the control and anosmia groups. Positive z-values (red color) indicate stronger structural covariance in the anosmia group compared to the control group. Conversely, negative z-values (blue color) signify weaker structural covariance in the anosmia group relative to the control group. Asterisks in panel C indicate significant between-group differences after FDR correction.

Table 2.

Group-level correlation coefficients reflecting conectivity strength among key brain regions in the control and anosmia groups

Key brain regions	Control (n=30)	Anosmia (n=22)
Amygdala - Entorhinal cortex	0.437	-0.195
Orbitofrontal cortex - Amygdala	-0.006	0.597
Orbitofrontal cortex - Entorhinal cortex	0.056	0.071

Partial correlation analysis was performed within both the control and anomia groups to assess connectivity strength among the amygdala, entorhinal cortex, and orbitofrontal cortex. Intracranial volumes were incorporated into the model as a relevant covariate.

Table 3.

Comparisons of connectivity strength between key brain regions involved in olfactory processing acorss the control and anosmia groups

Key brain regions	z	FDR-corrected p
Amygdala - Entorhinal cortex	-2.22	0.039
Orbitofrontal cortex - Amygdala	2.32	0.039
Orbitofrontal cortex - Entorhinal cortex	0.04	0.968

Z-test statistics were employed to compare the group-level correlation coefficients between the anosmia and control groups. P values were corrected for multiple comparisons using the False Discovery Rate (FDR) method.

In the anosmia group, the connectivity strength between the amygdala and entorhinal cortex was found to be weaker compared to that in the control group (z=-2.22, FDR-corrected p=0.039). Conversely, a higher connectivity strength between the orbitofrontal cortex and amygdala was observed in the anosmia group than in the control group (z=2.32, FDR-corrected p=0.039). However, there was no significant difference in the connectivity strength between the orbitofrontal cortex and entorhinal cortex across the groups (z=0.04, FDR-corrected p=0.968).

DISCUSSION

In this study, we employed structural covariance analysis to investigate brain reorganization patterns in key areas involved in olfactory processing: the amygdala, entorhinal cortex, and orbitofrontal cortex. We observed the reduced connection between regions of the primary olfactory network, specifically between the amygdala and entorhinal cortex, in the anosmia group. In contrast, the structural connections between the orbitofrontal cortex—a critical component of the secondary olfactory network—and the amygdala were enhanced in the anosmia group compared to the control group.

Our findings of reduced structural integration among key regions within the primary olfactory network align with prior studies that reported atrophic changes in the primary olfactory cortex associated with anosmia [27-29]. Brain regions within this network are crucial for detecting and initial processing of olfactory information [30, 31]. Considering that acquired anosmia frequently results from damage to the primary olfactory network due to traumatic brain injury or upper respiratory tract infection [32-34], the structural disintegration observed in our study reflects a major pathological mechanism of anosmia.

Furthermore, regions of the primary olfactory cortex, such as the amygdala and entorhinal cortex, are crucial for linking odors to emotional responses and memories [35, 36]. This integration is essential for assigning emotional values to smells and memory formation, enabling the recall and recognition of odors [37, 38]. Moreover, damage to the entorhinal cortex can significantly impair the ability to learn and retain new olfactory information [39]. From a clinical perspective, the structural disintegration observed in olfactory-related brain regions, including the amygdala and entorhinal cortex, may explain why anosmia can manifest as an early symptom of cognitive decline in conditions like dementia or Alzheimer’s disease [7, 27, 40]. Furthermore, our observed connectivity reduction has clinical implications for patients with anosmia, potentially affecting cognitive and memory functions [41, 42]. Taken together, these findings highlight the importance of a comprehensive approach in understanding and addressing the multifaceted consequences of olfactory loss.

Another significant finding of this study is the enhanced structural connections between the orbitofrontal cortex and amygdala in the anosmia group compared to the control group. The orbitofrontal cortex, as a higher-order olfactory region, plays a crucial role in integrating olfactory information with inputs from other sensory modalities [43-46]. Specifically, the orbitofrontal cortex can receive inputs from taste, visual, and tactile senses, while also maintaining reciprocal connections with primary olfactory network regions. This multimodal sensory integration enables a comprehensive and unified perception of the environment [43]. Furthermore, this process is crucial not only for recognizing and responding to complex stimuli but also for the cognitive and emotional aspects of olfaction [44].

Therefore, our observations of enhanced structural connections between the orbitofrontal cortex and amygdala, a major component of the primary olfactory network, suggest a potential compensatory adaptation of the brain to insufficient olfactory input. Specifically, these alterations may enhance the processing of residual olfactory cues by strengthening the integration of alternative sensory modalities in patients with anosmia. Consistent with prior findings that brain adaptations to visual deficits can improve auditory processing capabilities [47, 48], our results of strengthened structural integration between primary and secondary olfactory network regions might reflect cross-modal neuroplastic changes in response to the loss of olfactory sensation [49, 50].

It is important to note that our analysis revealed no significant alterations in structural covariance between the orbitofrontal cortex and entorhinal cortex across the study groups. Considering the average disease duration of approximately 5.7 years in our patient group, our results might suggest the reversible changes or neuronal reorganization in response to the loss of olfactory sensation, rather than irreversible neuronal damage within the brain’s olfactory network. This finding holds potential clinical relevance, providing brain anatomical insights that could guide targeted interventions for anosmia patients. However, further longitudinal research investigating both acute and chronic anosmia cases is necessary to confirm and refine our findings.

Several limitations should be considered when interpreting the results. Notably, this study did not examine the structural covariance of the piriform cortex, a crucial part of the primary olfactory cortex, with other olfactory-related brain regions. This was partly due to the limitations in segmenting this region using the FreeSurfer framework. Future studies should investigate a broader range of olfactory circuit regions to validate and expand upon our findings. The relatively small sample size and the cross-sectional design of our study could potentially limit the generalizability and inferential power of our findings. Large-scale, longitudinal studies on anosmia patients are needed to validate our results and more comprehensively investigate the reversibility of structural covariance alterations.

The structural covariance analysis of brain regional volumes does not yield individual measures for structural connections, as it relies on the correlation coefficient across a group of subjects [51]. Consequently, we could not directly investigate the relationship between structural connections of olfactory-related brain regions and other clinical measures. Future studies utilizing diffusion tensor imaging could offer individual-level assessments of structural connectivity within the olfactory network. This would provide further insights and potentially support our findings suggesting a compensatory role of enhanced structural connections in anosmia patients.

In conclusion, our study reveals structural reorganization of the brain associated with anosmia. We demonstrate diminished connectivity between the amygdala and entorhinal cortex highlighting the disruptive impact of anosmia on the olfactory processing network. Additionally, the enhanced connectivity between the orbitofrontal cortex and amygdala suggests potential compensatory mechanisms in response to olfactory loss. These findings indicate a potential restructuring of brain regions, especially involving the amygdala, entorhinal cortex, and orbitofrontal cortex, as a response to olfactory deficits. Future research is necessary to explore the functional implications of these structural changes and their potential role in developing clinical interventions for anosmia.