Introduction
Different performances between sexes on various cognitive skills, including executive control (Gaillard et al.
2021a), language (Wegesin
1998), spatial thinking (Zancada-Menendez et al.
2016; Gur and Gur
2017), emotional processing (Whittle et al.
2011) and social cognition (Gur and Gur
2017), have been explained in the last decades on the base of between-sex dissimilarities in brain function (Gaillard et al.
2021b; Whittle et al.
2011; Stevens and Hamann
2012; Volf et al.
2010; Filippi et al.
2013; Gur and Gur
2017). This evidence has supported the hypothesis that males and females engage different strategies depending on task demands related to different neural networks and may indicate adaptive diversity and complementarity between the sexes (Gur and Gur
2017). Furthermore, this aspect may represent a critical step toward a better understanding of sex-related pathological processes underlying most neurologic and psychiatric disorders characterized by sexual dimorphism (Lawrence et al.
2020; De Micco et al.
2019; Bakeberg et al.
2021). This, in turn, would help develop in the future personalized treatments for different neurological disorders tailored for male and female populations, respectively.
An appealing unbiased strategy to assess activity differences between sexes is the evaluation of resting-state (RS) functional connectivity (FC). In recent years, FC fMRI approaches found several brain regions whose spontaneous low-frequency fluctuations (< 0.1 Hz) of the blood oxygen level-dependent (BOLD) signal registered during RS correlate with each other. Those regions are believed to be functionally connected (Biswal et al.
2010; Greicius et al.
2003; van den Heuvel and Hulshoff Pol
2010). Based on their functional and/or anatomical overlap with well-known large-scale functional networks, distinct resting-state functional networks (RSNs) have been identified (Damoiseaux et al.
2006; Seeley et al.
2007), representing the brain’s function beyond explicit tasks and, consequently, the intrinsic functional architecture of the human brain (Smith et al.
2009). Using independent component analysis (ICA), several studies have demonstrated the occurrence of high temporal correlations between spatially distinct but functionally related brain regions, resembling specific neuroanatomical networks, which characterize the RSNs (Damoiseaux et al.
2006; Tedeschi and Esposito
2012). Furthermore, the spectral composition of RS-fMRI signals, as quantified by the fractional amplitude of low-frequency fluctuations (fALFF) (Zou et al.
2008) across four canonical frequency sub-bands (slow-5, 0.01–0.027 Hz; slow-4, 0.027–0.073 Hz; slow-3, 0.073–0.198 Hz; slow-2, 0.198–0.25 Hz) has been also considered to further characterize the BOLD time-courses within the RSNs under pathological (neurodegenerative) conditions (Esposito et al.
2013).
Recent studies explored the influence of sex on the RSNs, revealing that men showed stronger connectivity in parieto-temporal regions, and within cognitive and sensory networks, while women displayed stronger connectivity in fronto-temporo-cerebellar regions, and within attention and memory-related networks (Filippi et al.
2013). Moreover, RS-fMRI and graph theoretical approaches, used to investigate the hemisphere- and gender-related differences in brain functional networks in 86 young, healthy, right-handed adults, showed that males tended to be more locally efficient in the right hemispheric networks, but females tended to be more locally efficient in the left hemispheric networks (Tian et al.
2011).
Among neurodegenerative disorders characterized by sexual dimorphism (Pinares-Garcia et al.
2018), amyotrophic lateral sclerosis (ALS), a devastating motor neuron disease causing the progressive impairment of motor function (speech, swallowing, limb, respiration), has been shown to be strongly influenced by sex, together with age and genetic variations, in phenotypic manifestation (i.e. onset, phenotype and early progression of motor and cognitive symptoms) (Chiò et al.
2017,
2020; Rooney et al.,
2017a,
b; Trojsi et al.
2019). Sex has been reported as an independent factor involved in the development of ALS, with the higher risk of exhibiting the disease in men (Chiò et al.
2017), especially in case of flail limbs and respiratory phenotypes, with a trend toward a higher frequency in older age (Chiò et al.
2020). Moreover, the role of male sex in clinical presentation and prognosis of ALS patients carrying repeat expansions in chromosome 9 open reading frame 72 (
C9orf72) has been also explored (Rooney et al.
2017a; Trojsi et al.
2019). Nevertheless, while there is consolidated clinical and neurobiological evidence in favor of sexual dimorphism in ALS, from the neuroimaging point of view, only one MRI study investigated sex differences in the structural patterns of cortical and subcortical pathology in ALS (Bede et al.
2014). In particular, Bede et al. (
2014) revealed major between-sex differences in diffusion tensor imaging (DTI) and cortical thickness measures in fronto-temporal and cerebellar regions.
On this background, in order to start shedding some light on the still unclear functional MRI brain correlates of sex differences in ALS, we performed a RS-fMRI analysis with ICA, quantifying component-level fALFF, in a cohort of 45 patients with classical ALS compared to 30 healthy subjects, stratifying both cohorts for sex. In addition, we performed a whole-brain voxel-based morphometry (VBM) analysis to assess gray matter (GM) volume changes associated to sex differences within the same cohorts. We expected to find gender-related RS-fMRI and VBM patterns of brain damage useful to better characterize the differences in brain function and structure between sexes in ALS patients.
Discussion
This MRI study investigated, for the first time, the sexual dimorphism of brain RS-fMRI and VBM measures in ALS in comparison to health condition. Our findings, although confirming the evidence of a pattern of widespread brain RS-fMRI abnormalities extended to both motor and extra-motor RSNs, such as SMN, FPNs and SLN, in both female and male patients compared to sex-matched HCs, also revealed reduced FC in the DMN in the right middle frontal gyrus and in the precuneus and GM atrophy in the right lateral occipital cortex selectively in men with ALS compared to women with ALS. Conversely, fALFF analysis revealed an increasing trend of fALFF in the DMN in the slow-5 band in male ALS patients in comparison to both female ALS patients and male HCs.
While structural dimorphism in the human brain is well-described (Delvecchio et al.
2021; Seitz et al.
2021; van Eijk et al.
2020a,
b), controversy exists regarding the existence and degree of sex-related differences in brain function. In healthy conditions, several studies have described sex-related differences in fronto-parietal, cingulo-opercular and temporal connections in typically developing adults (Biswal et al.
2010; Zuo et al.
2010b) or individual networks such as the SLN and DMN in healthy aging (Jamadar et al.
2019). A key finding of a recent study by de Lacy et al. (
2019), who investigated 650 young adults matched for age and sex to determine the degree of sexual dimorphism in RSNs, was that most intrinsic networks exhibit significant sex-related effects, with both females versus males and males versus females effects usually found within the same network showing larger effects in women, although spreading across greater network territory in men. To note, sexual dimorphism was particularly common in task-positive control networks and in the DMN (de Lacy et al.
2019). In agreement with these findings and others that have found sex-related differences in the DMN in healthy conditions (Biswal et al.
2010; Zuo et al.
2010b), our results revealed that the function of the default mode system may be most strongly influenced by sex in ALS in comparison to other network types. In particular, between-sex comparisons revealed decreased FC in the right middle frontal gyrus and in the precuneus of the DMN only in affected men compared to affected women. Intriguingly, with regard to the posterior component of the DMN, affected men showed decreased FC in the posterior cingulate cortex when compared to unaffected men, while affected women showed an increased FC in the precuneus when compared to unaffected women. Taken together, these findings may corroborate the hypothesis that abnormal modulation of FC in the DMN may be a fingerprint of ALS-related alterations of brain FC, in agreement with previous RS-fMRI findings (Agosta et al.
2013; Trojsi et al.
2015b), additionally outlining that decreased FC in the posterior portion of the DMN is more characteristic of the male sex, recognized as a risk factor for early onset and respiratory and flail limb phenotypes of ALS (Chiò et al.
2020). Moreover, the evidence of adaptive diversity and complementarity of brain FC patterns between the sexes, revealed for the healthy condition (Gur and Gur
2017), seems to be confirmed also in ALS patients. On the other hand, a wide variety of human neurological and neuropsychiatric disorders showing sex-related differences in incidence, prevalence and severity has been associated with between-sex differences in FC modulation within DMN (Mohan et al.
2016).
To note, while ICA of RS-fMRI signals showed decreased FC in the right middle frontal gyrus and in the precuneus of the DMN in affected men compared to affected women, fALFF analysis revealed an increasing trend of fALFF in the DMN in the slow-5 band in male ALS patients compared to both women with ALS and healthy men. These findings represent further evidence that multiple, complementary analytical approaches are valuable for obtaining a more comprehensive characterization of RSNs alterations, from both spatial distribution and spectral composition point of views, as also revealed in previous RS-fMRI analyses of neurodegenerative, psychiatric or painful conditions (De Micco et al.
2019; Wolf et al.
2020; Flodin et al.
2014). In particular, we did not only use the ICA map representing the spatial distribution of the “within-network” functional connectivity, but we also used the ICA time-course representing the common signal fluctuations within this network (i.e., those signal fluctuations that are shared among more active regions). Now, the ICA time-course was further characterized in terms of the fALFF in multiple canonical frequency sub-bands (slow-5, slow-4, slow-3, slow-2). Moreover, by comparing network fALFF between groups across sub-bands, we interrogated the relative contribution of specific sub-bands to the previously mentioned common network time-course and, despite the reduction of the functional connectivity of one specific region to the entire network (e.g., DMN or FPN), the same network (as a whole) showed a reconfiguration of the spectral power of the functional connectivity time-course in terms of a trend toward higher contribution in a specific sub-band (slow-5) in male ALS patients compared to both women with ALS and healthy men. This might be interpreted as a compensatory phenomenon because the reduced contribution from one (spatially localized) region to the network is probably compensated by a relative higher contribution from other (spatially distributed) temporal sources that end up to affect the signal in a lower frequency sub-band (slow-5).
Amplitudes of low frequencies, evaluated by fALFF analysis have been proven to vary according to gender in healthy subjects (Lopez-Larson et al.
2011). Little evidence regarding low-frequency amplitudes have been reported in ALS: Ma et al. (
2016) revealed widespread fALFF changes in slow-4 and slow-5 bands and, more recently, Bueno et al. (
2019), who performed a whole-brain fALFF and regional homogeneity (ReHo) analyses, described widespread decreased fALFF and ReHo in several motor and sensory regions as well as cingulate, temporal, parietal and occipital areas in ALS patients compared to HCs (Bueno et al.
2019). These previous findings suggested abnormal neural activity in key nodes of SMN, DMN, FPN and SLN. In this regard, our result of an increasing trend of fALFF in DMN and R-FPN in the slow-5 band in male patients compared to female ALS patients recalls previous evidence of increased fALFF in middle and superior frontal gyrus in an ALS cohort compared to HCs, as reported by Ma et al. (Ma et al.
2016). Increased fALFF in several areas of RSNs in ALS patients has been speculatively attributed to activation of compensatory mechanisms in response to the neurodegenerative process (Douaud et al.
2011; Tedeschi et al.
2012; Agosta et al.
2013) or to increased inhibitory function attributed to “interneuronopathy” and reactive astrocytosis (Turner et al.
2012; Do-Ha et al.
2018). Probably, these mechanisms could substantially underlie cortical hyperexcitability and, consequently, increased FC revealed in some areas of brain RSNs in ALS patients compared to HCs (Douaud et al.
2011; Tedeschi et al.
2012; Agosta et al.
2013).
With regard to the brain functional abnormalities distinctive of the male population, decreased FC in the post-central gyri (SMN), in the right inferior parietal lobule (R-FPN) and in the ACC (SLN) were observed in our analysis in both ALS and healthy men compared, respectively, to affected and healthy women. On the other hand, ALS males showed increased FC in the right and left anterior insular cortices and decreased FC in the ACC in the SLN when compared to healthy men. While in our analysis the decreased FC in the ACC in the SLN was observed in both male and female patients compared to sex-matched HCs, in accordance with previous RS-fMRI findings from other cohorts of ALS patients compared to HCs (Trojsi et al.
2015b; Bueno et al.
2019), increased FC in the right and left anterior insular cortices, identified in our analysis only in men with ALS compared to healthy men, has been recently revealed in fast progressing phenotype of ALS in comparison to slow progressing phenotype (Trojsi et al.
2021). Consequently, the increased FC in insular cortices might represent a functional marker of poorer prognosis in male patients. To support and corroborate this hypothesis, in the future, further longitudinal functional and structural MRI analyses in cohorts of genotypically and phenotypically well-characterized men and women with ALS could be performed.
VBM between-sex analysis showed GM atrophy in the right lateral occipital cortex only in men with ALS in comparison to affected women. These results were in accordance with findings from cortical thickness analysis by Bede et al. (
2014), who revealed, accounting for diagnosis, a trend of higher age-adjusted cortical thickness in the right parieto-occipital and left mid-frontal regions in females, while males demonstrated higher cortical thickness in the left lingual and left superior temporal regions. Moreover, our findings may explain the inconsistency about occipital lobe involvement in gender-mixed cohorts of ALS: some studies (Kassubek et al.
2005) reported occipital lobe involvement, whereas others (Bede et al.
2016) revealed it as a less affected brain region in ALS. On the other hand, Delvecchio et al. (
2021) revealed non overlapping age-related, between-sex GM changes across post-adolescence in multiple cortical areas, including mid-occipital cortices and left inferior temporal gyrus. These results together with our observation of a lateral occipital brain damage in male patients may suggest a sex-related effect on cortical atrophy in ALS recalling that observed in ALS patients carrying
C9orf72 repeat expansions in comparison to shorter disease duration (sporadic) patients (Agosta et al.
2017) and that induced by age in healthy men compared to women (Delvecchio et al.
2021).
On the base of our findings and given the emerging evidence of gender differences in ALS, imaging studies focusing on ALS should invariably include sex as a covariate in their models, even if the groups are matched for gender. Furthermore, in the light of the need of a precision medicine approach, aimed at optimizing and individualizing treatment to the molecular drivers of an individual’s disease and beginning to be considered also in ALS, sex should arise as a variable to be considered in interpretation of data from clinical trials performed in cohorts of ALS patients. In fact, the observed trends of between-sex differences, derived from our and previous MRI analyses, might emerge from this interpretation. Moreover, the interaction between genetic, demographic, environmental, and lifestyle risk factors seems to underlie the pathological process inducing ALS, which would comprise deficits in multiple pathways, reflecting a “multistep” model of disease consistent with a six-step process (Chiò et al.
2018). Among demographic factors, sex has been reported as an independent factor influencing the development of ALS, being higher the risk of developing the disease in men with a trend toward higher frequency in older age (Chiò et al.
2020).
There are limitations to our study. First, the studied sample was relatively small. Second, we were not able to perform a longitudinal MRI study, because of the lacking consent of most patients to repeat the MRI exam. Third, our studied cohort had a mean age comprised between 50 and 60 years and, considering the proposed protective effect of female hormones, a gender study of younger patients might potentially reveal further differences. Moreover, we excluded from our analysis more disabled patients hindered to undergo an MRI exam because of diaphragmatic weakness and therefore we cannot exclude that this might have biased the cohort towards less severely disabled ALS patients. Finally, we did not record continuous blood oxygen saturation via oximetry during MRI scan.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.