Functional status is a patient-centered outcome that quantifies the burden of a disease and its impact on individuals' participation in society. Among the elderly, the trajectory of functional status is complex and influenced by several factors (Dhamoon et al., 2014). In particular, MRI measures of subclinical vascular disease are emerging as important predictors of outcomes (Dhamoon et al., 2017). White matter hyperintensities (WMH) in the brain are likely caused by traditional vascular risk factors (Hofer et al., 2015), and they have been associated with subsequent stroke [hazard ratio (HR) 1.86, 95% confidence interval (CI) 1.35–2.56; (Ntaios et al., 2015)] mortality [HR 1.64 (1.37, 1.97) for WMH grade 3 vs. 0 or 1; (Kuller et al., 2007)], cognitive impairment (drop of 2.4 vs. 0.9 points on the Modified Mini-Mental State Examination; Longstreth et al., 2002; Hachinski, 2008) and functional impairment (with 2–3-fold greater impairment with greater WMH; Pohjasvaara et al., 2007; Baune et al., 2009; Poggesi et al., 2011).

We previously showed that overall brain white matter hyperintensity volume (WMHV) was associated with accelerated long-term functional decline, independently of vascular events (Dhamoon et al., 2015). However, it was unclear whether asymmetry of WMHV burden in particular brain regions is more predictive of decline, independently of overall WMHV. In the normal brain, there is asymmetric localization of specific brain functions such as, language, but reliable structural asymmetries have not been found. With individual patients, neurologists are able to anticipate the impact of an asymmetric acute vascular brain lesion on functional status (Yaghi and Elkind, 2015), but the influence of asymmetrical subclinical white matter changes on functional status among community-dwelling healthy individuals is uncertain. Asymmetry of WMHV has not been previously examined as an independent predictor of functional status, but it has the potential to independently explain functional trajectories, as brain networks involved in functional tasks are bilaterally distributed and can be impaired not only by total lesion burden but also asymmetry of burden (Volz et al., 2017).

We analyzed repeated measures of functional status in a large population-based imaging study, and examined the differential effect on functional trajectories of left vs. right WMHV in different brain regions; we will refer to this as examining the effect of asymmetry on functional trajectories. We hypothesized that greater WMHV asymmetry in particular brain regions would be more predictive of functional decline, independently of baseline confounders, vascular events occurring during follow-up, and total brain WMHV.

The Northern Manhattan Study (NOMAS) prospective cohort includes an MRI substudy (as previously described by Elkind et al., 2006) that began in 2003 and included 1,290 individuals: (1) aged ≥50 years, (2) without MRI contraindications, (3) without clinical stroke, and (4) able to provide signed informed consent. Imaging was performed on a 1.5 T MRI system (Philips Medical Systems, Best, Netherlands), and included axial T1, axial T2, and Fluid Attenuated Inversion Recovery (FLAIR) sequences.

We developed tailored protocols using tools from the FSL software package (http://www.fmrib.ox.ac.uk/fsl) to automatically measure total WMHV and regional distributions over 14 brain regions [brainstem, cerebellum, and bilateral frontal, occipital, temporal, and parietal lobes, and bilateral anterior and posterior periventricular white matter (PVWM), defined as within 1 cm of the lateral ventricles; (DeCarli et al., 2005)]. The 14 regions selected for this study were chosen in an a priori fashion based upon standard neurological divisions into brain lobes. We did not use all available brain regions, greater than these 14, because the inclusion of too many regions reduces statistical power (by increasing the standard deviation) and reduces reliability especially in subjects who have an already low load of WMH volume. T2 FLAIR images were used for white matter segmentation. The skull region was removed from images using the FSL-Brain Extract Tool (Smith, 2002), and two-class segmentation of the whole brain was performed, into cerebrospinal fluid and brain tissue using the FSL-FAST segmentation algorithm after correction for image non-uniformity (Zhang et al., 2001). The mean (μ) and standard deviation (σ) of intensity values of the brain tissue voxels were derived, and WMH voxels were defined with intensity >(μ+3.5^*σ). Finally, regional WMHV were calculated from T2-FLAIR derived images using the Montreal Neurological Institute structural atlas as a reference by nonlinearly registering the subject images to the atlas template and mapping the regions delineated on the atlas to the subject images using FSL-FNIRT (Non-Linear Image Registration) tool (Andersson et al., 2008). Total WMHV was the sum of volumes in each of the 14 sub-regions. The reliability of the WMH volumetry method was previously established using repeated measurements of the same subjects (Alperin et al., 2014). Total cranial volume (TCV) constituted the sum of whole brain volume voxels from the T1 segmentation process. Our segmentation methods differ from similar published methods (Iorio et al., 2013) because we calculated regional distribution, and optimized our thresholds for selection to minimize the need for manual editing. An example of segmentation images is provided in Figure 1.

Eighty-five participants were excluded due to: lack of FLAIR data, image artifacts in FLAIR data, including motion artifact, failure of the registration method due to very large ventricles or lesions/tumors, and lesions that appeared hyperintense outside the white matter. Columbia University and University of Miami IRBs approved the study and all participants provided written informed consent.

Table 1 shows distributions of baseline variables. Mean follow-up time, from MRI to last follow-up assessment, was 7.3 years (SD 2.1). Of 1,136 participants with BI of 95 or 100 at baseline and >1 follow-up, 690 (61.7%) experienced a decline in BI. Mean WMHV (as % TCV) by region are reported in Table 1, and the majority of WMHV burden involved the anterior and posterior periventricular regions. When the proportional right-left difference of regional WMHV was examined (Table 2), more individuals had right > left predominance for frontal, occipital, and parietal lobes and anterior and posterior periventricular regions, with the greatest proportional difference in the occipital lobe, followed by parietal and temporal lobes.

Figure 3 shows results from fully adjusted models testing the association between regional right-left WMHV asymmetry and (1) baseline functional status (Figure 3A, conceptually depicted in Figure 2A) and (2) additional annual change in functional status (Figure 3B, conceptually depicted in Figure 2B). Greater asymmetry, with greater WMHV on the right compared to the left, was associated with lower baseline functional status in the parietal (−8.46 BI points per unit WMHV difference, 95% CI −3.07, −13.86) and temporal lobes (−2.48, 95% CI −1.04, −3.93). There was a trend for an association between greater WMHV asymmetry (right > left) in the entire brain and lower baseline BI score that did not reach significance (−8.70, 95% CI 0.62, −18.03).

When the effect of WMHV asymmetry on slope of functional status (Figure 2B) was examined, the estimated decline in functional status ranged from −0.97 to −1.03 BI points per year in all models. Asymmetry in the parietal region was associated with steeper annual decline: one unit greater WMHV burden in the left parietal lobe compared to the right was associated with an additional decline of −1.09 BI points per year (95% CI −1.89, −0.28, Figure 3B).

When mobility and non-mobility domains of the BI were examined as separate outcomes (Figure 4), similar patterns as above persisted for each domain. Namely, there were associations of lower baseline BI score, in both mobility and non-mobility domains, with greater WMHV asymmetry (right > left) in the parietal and temporal lobes (Figure 4A). There was a trend for an association with WMHV asymmetry in the entire brain for each domain. When the slope of mobility and non-mobility domains of functional status was examined (Figure 4B), parietal WMHV asymmetry (left > right) was associated with steeper annual decline for both domains. In addition, asymmetry in the temporal lobe (left > right) was associated with steeper annual decline in the mobility domain: one unit greater WMHV burden on the left compared to the right was associated with an additional decline of −0.18 BI points per year (95% CI −0.35, −0.01).

We examined the differential effect, on functional trajectories, of left vs. right WMHV in different brain regions, which we term asymmetry. We found that greater asymmetry of WMHV in specific brain regions was associated with functional trajectories, independently of confounders, stroke and MI occurring during follow-up, and total brain WMHV. In particular, greater WMHV burden in the right parietal and temporal lobes, compared to the left, was associated with lower baseline functional status. Also, greater WMHV burden in the left parietal lobe, compared to the right, was associated with steeper annual functional decline, with a doubling of decline per unit right-left WMHV difference. Results were consistent when mobility and non-mobility functional domains were examined separately.

WMH in the brain are likely caused by ischemic processes (Conklin et al., 2014), but other presumed mechanisms include blood brain barrier disruption, inflammatory processes, ependymal loss, and venular dysfunction (Black et al., 2009). Rarer causes of white matter changes include cerebral amyloid angiopathy and Cerebral Autosomal Dominant Arteriopathy with Subcortical Infarcts and Leukoencephalopathy. The resulting WMH may disrupt cholinergic pathways involving the cortex, leading to dysfunction in executive, memory, and visuospatial domains that negatively impact functional status (Kim et al., 2011). However, these mechanisms do not explain why asymmetric WMHV may be associated with worse function. Although overall and regional WMHV have been tested as predictors of functional outcomes, the independent effect of WMHV asymmetry has been understudied.

Asymmetry is a fundamental aspect of normal brain function, with lateralization of language function, motor dominance, and perception (Toga and Thompson, 2003). However, in a healthy brain, macrostructural asymmetries have not been reliably found, either in gray or white matter volume, or sulcal or gyral sizes; the functional basis of the asymmetry may reside in lateralized synaptic networks or microstructural white matter tract differences (Büchel et al., 2004; de Schotten et al., 2011; Thiebaut de Schotten et al., 2011). It is conceivable that asymmetrical lesion burden or pathology in diverse neurological conditions may be associated with functional status, independently of total lesion burden. There are several possible justifications for such an association. Lateral injury may disrupt excitatory or inhibitory interhemispheric interactions that could impair performance of the complex tasks required for ADLs (Takeuchi et al., 2012). For example, asymmetry of frontal white matter fractional anisotropy measured by diffusion tensor imaging was associated with worse executive functioning (Yin et al., 2013). Among 11 individuals with recent ischemic stroke (Graziadio et al., 2012), the degree of asymmetry in corticospinal tract activity, measured by electroencephalography and electromyography, was independently associated with poorer recovery. It appears that symmetry of brain function and structure may be important for optimal functioning, and asymmetric WMHV burden is associated with worse functional trajectories, even when adjusting for total brain WMHV, as we have found. Furthermore, although asymmetry of vascular anatomy is common, especially in the transverse sinuses (Alper et al., 2004), the relationship between this asymmetry and white matter asymmetry has yet to be explored.

We found that greater WMHV burden in the right parietal and temporal lobes, compared to the left, was associated with lower baseline functional status. Right-sided clinical stroke has been previously associated with accelerated functional decline over the long term, independently of age, vascular risk factors, and stroke severity (Dhamoon et al., 2009). Subclinical WMHV may similarly cause impaired cross-sectional or baseline functional status. Also, greater WMHV burden in the left parietal lobe, compared to the right, was associated with steeper annual functional decline, with a doubling of decline per unit right-left WMHV difference. Disruption of pathways involved in parietal association cortices may impact the cognitive and visuospatial aspects of functional status and result in accelerated long-term decline. These patterns of association were consistent when mobility and non-mobility functional domains were examined separately, suggesting that asymmetric white matter injury not only affects motor pathways but also those that are involved in executive functioning, visuospatial perception, and ordering of complex tasks.

We previously showed, in this cohort, that periventricular WMHV was uniquely selected, among all tested brain regions, with accelerated functional decline over time (Dhamoon et al., 2017). Indeed, multiple studies have shown that periventricular WMHV may develop earlier, and with greater burdens, than in other brain regions (Wakefield et al., 2010; Moscufo et al., 2011; Zheng et al., 2012). Greater periventricular WMHV has been associated with lower mobility and gait speed (Moscufo et al., 2011), physical decline (Zheng et al., 2012), and falls (Blahak et al., 2009). However, we found no association between periventricular WMHV asymmetry and functional trajectories, likely because periventricular WMHV accumulates symmetrically over time (Gunning-Dixon et al., 2009). Indeed, the smallest differences in right-left WMHV in our study were seen in periventricular regions. Periventricular WMH have been associated with impaired insight and different temperament profiles were associated with differences in the subcortical structures of the brain (Serafini et al., 2016). Those with lower insight had significantly more periventricular WMH when compared to patients with higher insight. Also, differences among temperament groups were associated with differences in WMH patterns (Serafini et al., 2011). Further study would clarify relationships between WMH patterns and factors that affect performance in ADLs.

Strengths of this study include the large population-based cohort, accurate assessment of events during follow-up, minimal loss to follow-up, the use of state-of-the-art imaging and regional measurement of WMHV, and repeated measures of functional outcomes that allow trajectory analysis. As a result of these strengths, this is the first study to our knowledge to examine asymmetrical WMHV and associations with functional trajectories. A limitation of this study is that, in the MRI substudy, participants were recruited from the prospective cohort and most often obtained MRI imaging during follow-up instead of at baseline. The NOMAS participants enrolled in the MRI cohort were able to return for follow up and undergo MR imaging, reflecting a healthy survivor bias, which may have reduced power to detect declines in functional status. Also, although the ascertainment of follow-up events is excellent in NOMAS, there is the chance that some events occurring during follow-up may not have been captured. Finally, 1.5 T MRI was used in this study, and there is the chance that using 3 T MRI would have resulted in greater resolution for WMHV and greater overall volumes measured.

In conclusion, we found that greater asymmetry of WMHV in particular brain regions was independently associated with worse functional trajectories over time. These findings suggest that not only overall burden of WMHV, but also degree of asymmetry of this burden, may be important independent prognostic signs of functional limitations. Further research will clarify whether asymmetric WMHV causes impairment by disrupting interhemispheric brain networks, over and above direct damage to motor and association pathways.

This study was carried out in accordance with the recommendations of the IRB of Columbia University and the University of Miami with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the IRB of Columbia University and the University of Miami.

MD: study design, interpretation, analysis, and writing. YC, AB, and NA: analysis and writing. RS: study design and writing. ME and CW: study design, analysis, and writing.