Most patients with late-life depression (LLD) have cognitive impairment, and at least one-third meet diagnostic criteria for mild cognitive impairment (MCI), a prodrome to Alzheimer’s dementia (AD) and other neurodegenerative diseases. However, the mechanisms linking LLD and MCI, and brain alterations underlying impaired cognition in LLD and LLD + MCI remain poorly understood.
To address this knowledge gap, we conducted a systematic review of studies of brain-cognition relationships in LLD or LLD + MCI to identify circuits underlying impaired cognition in LLD or LLD + MCI. We searched MEDLINE, PsycINFO, EMBASE, and Web of Science databases from inception through February 13, 2023. We included studies that assessed cognition in patients with LLD or LLD + MCI and acquired: (1) T1-weighted imaging (T1) measuring gray matter volumes or thickness; or (2) diffusion-weighted imaging (DWI) assessing white matter integrity. Due to the heterogeneity in studies, we only conducted a descriptive synthesis.
Our search identified 51 articles, resulting in 33 T1 studies, 17 DWI studies, and 1 study analyzing both T1 and DWI. Despite limitations, reviewed studies suggest that lower thickness or volume in the frontal and temporal regions and widespread lower white matter integrity are associated with impaired cognition in LLD. Lower white matter integrity in the posterior cingulate region (precuneus and corpus callosum sub-regions) was more associated with impairment executive function and processing speed than with memory.
Future studies should analyze larger samples of participants with various degrees of cognitive impairment and go beyond univariate statistical models to assess reliable brain-cognition relationships in LLD.
Late-life depression (LLD) is typically defined as major depressive disorder (MDD) occurring in adults 60 years or older. Estimates of the prevalence of LLD range from 2–5% among community-dwelling older adults, reaching up to 12% among hospitalized older adults . Most patients with LLD present with more cognitive impairment than age-matched non-depressed controls, and about one-third meet diagnostic criteria for mild cognitive impairment (MCI) . While cognitive deficits associated with LLD can involve any domain of cognition [3, 4], executive function and processing speed are considered to be the core cognitive deficits in LLD [5, 6] and may mediate deficits in other cognitive domains [3, 7, 8]. LLD is also associated with an increased risk of developing dementia of all causes and Alzheimer’s dementia (AD) in particular . Patients with both LLD and MCI (LLD + MCI) have two risk factors for developing AD and therefore could be at an increased risk. LLD can occur first later in life after the age of 60 and is then referred to as late-onset depression (LOD), or it can have occurred first early in life and recurred in late life and is then referred to as early-onset depression (EOD). Both LOD and EOD have been associated with an elevated risk for AD . Several meta-analyses have confirmed that EOD is a risk factor for all-cause dementia and AD [9, 11, 12]. Compared to EOD, LOD has been associated with more severe cognitive impairment [13, 14], and with white matter hyperintensities and other cerebrovascular abnormalities that are common in AD . This led some studies to suggest that LOD is an early neuropsychiatric manifestation of AD .
Some of the neurobiological mechanisms explaining the relationship between LLD, MCI, and AD have been summarized through two pathways. First, models of the vascular pathway suggest that cerebrovascular disease, and ischemic lesions in particular, lead to executive dysfunction. Second, models of the inflammation pathway suggest that elevated levels of stress hormones promote neurodegeneration, particularly hippocampal volume loss, thus leading to impaired episodic memory . Other mechanisms linking LLD and MCI include Alzheimer’s pathology (e.g., amyloid beta accumulation) in brain regions related to mood regulation serving as a contributor to depression [17, 18], and cerebral blood flow reductions in brain regions related to mood and cognitive symptoms . A comprehensive discussion of the mechanisms linking LLD and AD is beyond the scope of this review and is the topic of other reviews [16, 17, 19, 20].
Several neuroimaging studies have identified brain structural alterations associated with LLD or LLD + MCI that contribute to the risk of AD [21,22,23]. However, the literature exploring the association between structural abnormalities and cognitive impairment in LLD and LLD + MCI remains sparse. To address this gap, we conducted a systematic review of studies of brain-cognition relationships in LLD or LLD + MCI. We aimed to identify vulnerable circuits underlying impaired cognition in LLD or LLD + MCI and reveal risk mechanisms for AD. We focused on T1-weighted imaging studies of gray matter (GM) assessing brain volume and thickness and on diffusion-weighted imaging (DWI) studies assessing white matter (WM) measures such as fractional anisotropy (FA) and mean diffusivity (MD). In secondary analyses, we explored whether brain-cognition relationships found across all studies are also present in LLD + MCI, as well as early and late-onset depression subgroups.
This systematic review was conducted in accordance with the updated Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020)  and registered on the International Prospective Register of Systematic Reviews (PROSPERO: CRD42022292905). The search strategy and protocol were reviewed by a librarian at the Center for Addiction and Mental Health (CAMH), Toronto, Canada, prior to registration and screening.
Information sources and search strategy
A systematic review of the literature was conducted using MEDLINE, PsycINFO, EMBASE, and Web of Science electronic databases from inception through February 13, 2023. Comprehensive search strategies adapted for Medline, EMBASE, PsycINFO, and Web of Science are available in Supplementary Materials. In brief, our search strategy included Medical Subject Headings (MeSH) and keywords related to three broad search blocks: geriatric depression (age group and condition being studied), structural magnetic resonance imaging (MRI, methodology of interest), and cognition (primary outcome measure). In addition to searching for articles of interest, reference lists of relevant review articles were also searched for additional eligible studies.
Eligibility criteria and study selection
Studies were included if they met the following criteria: (1) published in a peer-reviewed journal in the English language; (2) participants were older adults aged 55 years or older OR had a mean age of at least 65 years; (3) in at least one group, participants were formally diagnosed with MDD according to criteria from the Diagnostic and Statistical Manual of Mental Disorders (DSM) or International Classification of Diseases (ICD) (regardless of whether they also have MCI); (4) study assessed cognitive performance in diagnostic groups using at least one cognitive measure; (5) study acquired one (or both) of the following structural MRI modalities: T1-weighted scans measuring GM structure (i.e., volume or thickness) or DWI scans measuring WM integrity; and (6) study reported the results of some analysis of the relationship between the imaging measures and cognitive performance.
Studies were excluded if they: (1) focused exclusively on dementia or MCI; (2) exclusively studied depressive symptoms (i.e., without a formal diagnosis of MDD); (3) in the MDD group, included participants diagnosed with a major neurological illness (e.g., stroke, Parkinson’s disease, epilepsy, multiple sclerosis, traumatic brain injury); (4) in the MDD group, included participants with psychotic depression (i.e., did not separate those with MDD with or without psychotic features); (5) reported on case studies or non-human subjects. Conference abstracts, commentaries, opinion pieces, letters to the editor, and reviews were also excluded. As we conducted the review, we added an additional exclusion criterion to exclude clinical trials including participants who received electroconvulsive therapy (ECT) less than 3 months from the time of cognitive testing, due to potential effects of recent ECT on cognitive performance . If ECT was administered more than 3 months prior to the testing date, the study was included. However, no study met such criteria and therefore all ECT studies were excluded.
In accordance with the PRISMA guidelines, studies identified through searching the electronic databases first underwent title and abstract screening by one independent reviewer (TM) to determine their relevance with respect to the population, condition, methodology, and outcomes of interest. After removing any duplicates, full-text review of studies included from the screening stage was conducted. Two independent reviewers (TM, NJA) conducted the full-text review, and disagreements were resolved by a third study team member (BHM). Screening and study selection were conducted using the Covidence reference management system.
Data from studies meeting the eligibility criteria were extracted and entered into a database including bibliographic information, study type, sample size, mean age of the groups, age of depression onset, depression status, diagnostic criteria, reported treatment, categorical results of all cognitive assessments (e.g., domains with impaired cognition), imaging modality, imaging analysis and processing approach, statistical analysis methods, regions/tracts assessed, and a summary of study findings. Information was recorded for the LLD group and other additional groups (i.e., MCI, LLD + MCI, healthy controls - HC).
Risk of bias assessment
Quality assessment of all studies included was completed using a modified version of the Newcastle-Ottawa Scale (NOS) reported elsewhere . In summary, points were allocated to each study and summed up to range from 0–8, with scores between 0-3 indicating poor quality; 4–5, moderate quality; and 6+, good quality. Among our 51 included studies, 30 were classified as good quality, 19 as moderate quality, and 2 as low quality (Supplementary Table 5).
Overview of study characteristics
Our search identified 4,077 eligible studies after the removal of duplicates. After title and abstract screening, 3,838 were excluded and 239 progressed to full-text screening with 188 studies excluded at this stage, yielding 51 studies with 33 studies using T1 imaging, 17 using DWI, and 1 using both (Fig. 1). Table 1 presents the characteristics of the T1 longitudinal studies and Table 2 presents the T1 cross-sectional studies, including the way data were analyzed: region of interest (ROI), voxel-based morphometry (VBM), or structural network analysis. Table 3 presents the characteristics of the DWI studies, including the way data were analyzed: ROI or voxel-wise analysis.
Of the 51 studies, 41 compared LLD and HC groups (Supplementary Table 5); 7 included only an LLD group (Supplementary Table 6); and 3 studies included additional comparison groups: two studies compared patients with LLD, MCI, or LLD + MCI, and HC and one study compared patients with LLD or LLD plus a memory deficit (not meeting criteria for MCI), and HC (Supplementary Table 7). Thirty-eight studies included only patients during a current depressive episode, 10 studies included only patients with remitted depression, and 3 included patients with current or remitted depression. Ten of the 51 studies also compared findings in patients with EOD versus LOD, and 9 included only patients with LOD.
T1 longitudinal studies
Of the 33 T1 studies, 7 used a longitudinal design in participants with LLD and an HC group or LLD only [26,27,28,29,30,31,32]. Four of these studies reported significant positive associations between the bilateral hippocampal volumes and changes in the Mini-Mental State Examination (MMSE) scores during follow-up periods lasting from 21 months to 4 years [27, 30,31,32]. This association was not present in the HC group [27, 32].
These studies also evaluated the relationship between GM volumes and specific cognitive domains including tests of processing speed , visuospatial memory , response inhibition and set-shifting , and verbal fluency and memory . Hou et al. (2012) showed that larger right hippocampal volume predicted improved performance on the Symbol Digit Modalities Test over a 21-month period . No significant longitudinal associations between baseline hippocampal volume and persistent cognitive decline in episodic memory, executive function, or processing speed were reported over 18 months .
T1 treatment studies
Two studies explored pre- and post-treatment associations between GM volume and cognition during a treatment trial [26, 29]. In the first trial, 17 participants with LLD were treated with citalopram . In this voxel-based morphometry (VBM) analysis, larger baseline gray matter volumes of frontal regions (including the right superior and middle frontal gyri) and the left fusiform gyrus were associated with improvement in verbal fluency following treatment . Larger baseline volumes of the left middle frontal gyrus, left inferior frontal gyrus, right superior temporal gyrus, right uncus, bilateral fusiform gyrus, right angular gyrus, and right lingual gyrus were also associated with improvement in verbal memory following treatment . Interestingly, smaller baseline gray matter volume of the bilateral precuneus and superior frontal gyrus were associated with improved verbal fluency and verbal memory, respectively, following treatment . The second trial of 26 participants treated with venlafaxine did not find significant associations between longitudinal changes in brain structure and cognition following treatment .
T1 cross-sectional studies
Global cognition: MMSE or CAMCOG only
Twelve cross-sectional studies assessed the relationship between GM volumes or thickness and global cognition measured by the MMSE. Of these 12 studies, 4 reported significant positive associations with medial temporal lobe (MTL) region volumes including the hippocampus and hippocampal-amygdala complex in LLD [33,34,35,36], but not in HC [33, 36]. In two studies using an ROI analysis in patients with LLD, GM volume of the left DLPFC was positively associated with MMSE scores and left/right frontal lobe volume was positively associated with the Cambridge Cognition Examination (CAMCOG) scores [37, 38]. The latter finding was also significant in the HC group . A vertex-wise analysis of cortical thickness in an LLD group also reported a positive association between the thickness of the superior frontal gyrus bilaterally and MMSE scores . Finally, 6 of the 12 studies reported no associations between any GM measure and impaired global cognition [30, 39,40,41,42,43]. Only half of the cross-sectional studies assessing the relationship between GM volumes or thickness and global cognition reported associations.
Executive functioning and processing speed
Six studies used a detailed cognitive battery to test for brain-cognition associations with executive dysfunction and slower information processing speed in LLD. Two of these 6 studies used a whole-brain approach and reported that lower cortical thickness of frontal-executive (e.g., DLPFC, rostral middle frontal) and corticolimbic regions (e.g., superior prefrontal and frontal cortices, anterior cingulate, precuneus) was significantly associated with executive dysfunction in LLD [44, 45]. Of note, Lim et al. (2012) LLD group included only patients with LOD. Another study using an ROI approach identified an association between lower GM volume in the corticolimbic region, including the gyrus rectus and the orbitofrontal cortex (OFC), with executive dysfunction. . The association between the gyrus rectus volume loss and executive dysfunction differed significantly between the LLD and HC groups, with a weaker association in the LLD group . An association of lower cortical thickness in the superior temporal gyrus with executive dysfunction was also identified in the whole-brain analyses by Shin et al. (2018). There were no correlations between any measure of subcortical volumes and executive functioning . Finally, one study found significant associations between lower whole-brain GM volume and slower processing speed in LLD, but not in HC  . However, no associations between whole brain GM volumes using VBM and executive function were reported by Yuan et al. (2008) . Generally, cross-sectional T1 studies of brain-cognition relationships implicate different frontal and temporal regions.
Learning and episodic, verbal, or visuospatial memory
Fifteen studies examined associations between GM volumes or thickness and cognitive performance on tests of learning and episodic, verbal, or visuospatial memory. Of those, eight used an ROI approach, six conducted whole-brain voxel-wise or vertex-wise analyses, and one conducted a structural network analysis. The evidence strongly supports the role of MTL regions in learning and memory in LLD: 3 of the 8 ROI studies [7, 49, 50] and 3 of the 6 voxel-wise studies [44, 51, 52] found significant positive correlations between smaller bilateral hippocampal volume and lower scores of tests assessing episodic, verbal, and visual memory. These associations were also significant in remitted LLD . One study also implicated the parahippocampal gyrus in performance on immediate verbal memory tests . Two ROI studies implicated the OFC in verbal and visual memory [53, 54] in LLD, but not in HC . In their LOD vs. HC study, Lim et al. (2012) concluded that lower cortical thickness of the left superior temporal, precuneus, entorhinal cortex, and isthmus cingulate, correlated with lower verbal memory scores . Additional regions included the fusiform, insula, precentral, and supramarginal . However, another VBM study in patients with remitted LOD reported a contradictory association between volume of the cingulate gyrus and episodic memory, whereby larger volumes were correlated with worse memory functioning .
Both cortical and subcortical abnormalities in frontal-executive regions were specifically associated with poor verbal memory. Volume loss in subcortical regions, including the putamen, thalamus, and anterior caudate, was significantly associated with worse verbal recall in LLD [44, 55] but not HC [44, 55]. Similarly, volume loss in the rostral middle frontal, medial frontal, and caudal anterior cingulate as well as lower cortical thickness of inferior temporal and inferior parietal cortex were also associated with deficits in verbal memory in LLD [44, 53, 56]. In addition to LLD or HC groups, one single study included aMCI and LLD+aMCI groups . There were no associations between GM volumes and episodic memory in any of the groups. However, the sample size was small with <20 participants in each case group.
Finally, Shin et al. (2018) used a structural network correlation analysis in 50 participants with untreated LLD. They reported that a sub-network with cores belonging to the left subcentral cortex, right precuneus, and the posterior ramus of the right lateral sulcus was correlated with verbal fluency scores (but not episodic memory) . Single univariate associations were also identified between lower cortical thickness of the fusiform or lateral occipital gyri and worse episodic memory function, which agreed with other studies .
T1-weighted MRI comparisons of EOD vs. LOD
While nine T1 studies included both EOD and LOD groups [33, 35,36,37, 42, 49, 51, 55, 57], only 4 of them compared them in terms of brain-cognition relationships (with all 4 studies including participants with current depression) [35,36,37, 49]. The results of three studies support associations between GM thickness or volume and poor cognition in LOD, but not EOD or HC [35, 36, 49]. In the fourth study, conversely, whole-brain and bilateral frontal lobe volumes were associated with global cognition in participants with EOD or HC but not in those with LOD .
Eighteen of the 51 studies used DWI to assess the relationships between WM tract integrity measures (fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), or axial diffusivity (AD)) and cognition in LLD. Six studies used voxel-wise TBSS analysis, 11 studies used an ROI approach (of which 3 used tractography) and 2 studies used network-based statistics (NBS) (Table 3). Two studies included additional groups: an aMCI and LLD+aMCI groups  or an additional LLD with memory deficits group . Eleven of the 18 DWI studies included participants with acute depression, 6 included participants with remitted depression, and one study included a mixed sample (i.e., remitted + current).
TBSS studies (voxel-wise analysis)
Executive functioning and processing speed
Six studies used a voxel-wise analysis to explore the association of whole-brain WM tract integrity with executive function and processing speed. Of those, 4 studies reported positive associations in the LLD group [7, 59,60,61], 1 study in the HC group only , and 1 study reported no associations in any group . The results of studies with similar analytic approaches that identified a significant relationship with WM integrity included associations with some specific cognitive tests but not with others. For example, in one study, decreased FA in a cluster including the superior longitudinal fasciculus (SLF), SLF-temporal, and right corticospinal tract in the LLD group correlated with lower scores on the initiation/perseveration subscale of the dementia rating scale (DRS), but not with the Stroop Color Word Interference (CWI) task . In a comparable study, lower FA in fronto-striatal-limbic tracts in the LLD group was correlated with the Stroop CWI Task . One other study in remitted LLD showed that lower FA of the anterior thalamic radiation and the uncinate fasciculus was correlated with executive dysfunction, and lower FA of the genu of the corpus callosum was correlated with impaired processing speed .
Learning and episodic, verbal, or visuospatial memory
Two of the 4 DWI studies that assessed the association between WM tract integrity and performance on learning and memory tasks in LLD identified significant associations [7, 63], while two did not [59, 61]. Poor tract integrity (i.e., lower FA) in corticolimbic circuitry  and the genu/body of the corpus callosum and fornix  were associated with poor episodic memory. Alves et al. (2012) found an association between lower FA of the right posterior cingulate cluster and poor verbal fluency and episodic memory in their whole sample (LLD + HC). No significant associations were reported in the HC group .
Region of interest analysis
Executive function and processing speed
Of the 11 studies using an ROI approach, 10 assessed the relationships between WM tract integrity and executive function or processing speed. Five focused on the associations between corticolimbic tract integrity (i.e., the cingulate bundle and uncinate fasciculus) and executive function or processing speed in LLD [62, 64,65,66,67], with 4 reporting positive associations [62, 65,66,67]. One study selected cingulum bundle fiber tracts connecting the PCC/precuneus to the dorsal ACC, for which lower FA was associated with executive dysfunction across the whole sample (LOD + HC) . This finding implicating the posterior cingulate bundles was replicated in another study of remitted LOD . Higher integrity of the corpus callosum (lower MD) was associated with faster processing speed , but this finding was not replicated in another study with a similar regional analysis , potentially due to the remission status in their mixed sample of remitting and non-remitting participants . Regions of interest found in other studies included the anterior and posterior commissure and the anterior thalamic radiation, in which lower tract integrity (low FA) was associated with executive dysfunction [69, 70]. Finally, no brain-cognition associations with white matter integrity were found in remitted LOD . Despite the heterogeneity in findings, there is an overall trend of an association between lower white matter tract integrity and worse cognition.
Learning and episodic, verbal, or visuospatial memory
Eight DWI studies examined the relationships between integrity of selected WM tracts with learning and memory in LLD [21, 64,65,66,67,68, 70, 71]. Six of these studies reported no significant brain-cognition associations [64, 66,67,68, 70, 71]. Only two studies reported significant associations [21, 65]. In the first, lower FA of the cingulate and uncinate fasciculus was associated with poor episodic memory and language; however, this correlation was weaker than that with executive function or processing speed . The second study reported associations between higher diffusivity (i.e., MD and RD), but not lower anisotropy, values of the cingulum-hippocampus tract and poor episodic memory across patients with LLD, aMCI, LLD+aMCI, and HC . However, they did not find any significant associations within each case group .
Additional approaches: network-based statistics and machine learning
Two DWI studies used probabilistic tractography and network-based analysis (NBS) to explore the relationship between WM connectivity and cognition in remitted LLD , LOD, or LOD + memory deficits . In the first study, less connectivity in a subnetwork comprising the left lingual gyrus, middle occipital gyrus, and fusiform gyrus was associated with slowed processing speed in the remitted LLD group, but not in HC . In the other study, altered connectivity in a global subnetwork spanning frontal, temporal, parietal, and subcortical regions in LOD  was associated with executive function, processed speed, and memory, with the highest correlation with processing speed .
A third DWI study used relevance vector regression (RVR), a multivariate machine learning approach, to predict the contribution of multimodal MRI features (i.e., T1 and DWI) to slowed processing speed in remitted LOD patients . Their model identified 26 GM volumes and 8 WM features (3 FA and 5 MD), whereby lower GM volumes and lower tract integrity predicted worse processing speed .
We conducted a systematic review of 51 MRI studies (33 T1 studies, 17 DWI studies, and 1 study analyzing both T1 and DWI) of brain-cognition relationships in patients with LLD. Three main findings emerge from these studies, including consistent gray matter regions and white matter circuitry implicated in cognitive impairment in LLD. First, overall, these studies support the role of altered corticolimbic circuitry (particularly in the hippocampus, precuneus, entorhinal cortex, and cingulate cortex) in deficits of learning and memory. DWI studies more consistently implicate the cingulate bundle and other posterior cingulate clusters including the corpus callosum sub-regions in executive dysfunction. Second, measures of WM integrity were more strongly correlated with executive dysfunction than with memory impairments. Third, more consistent brain-cognition relationships emerged for LOD than EOD.
Confidence in these results is tempered by methodological limitations in the literature and in our review. The relevant studies we identified were heterogeneous in terms of both participant profiles and imaging methods. For example, LLD participants varied in depression status, age of onset, or treatment status. Also, most studies excluded patients with LLD and comorbid MCI or dementia based on their MMSE scores. However, the MMSE lacks sensitivity to cognitive deficits in patients with LLD . Thus, cognitive impairment in patients with LLD in these studies may contribute to the heterogeneity of results. Similarly, our discussion of relationships between brain structure and global cognition is limited by the prevalent use of the MMSE as a measure of global cognition, which presents with major limitations (e.g., poor test-retest reliability) as an accurate measure of interindividual variability in cognition . Another limitation of the surveyed literature is the relatively small sample sizes, ranging from 12 to 99 participants (with the exception of one study ). These findings should be replicated in future studies with larger sample sizes featuring hundreds of participants to identify replicable brain-behavior associations [76, 77]. Moreover, we identified only one T1 study  and one DWI study that compared LLD to MCI or LLD + MCI; more studies should compare participants with LLD and varying cognitive profiles (including LLD + MCI). Findings from those two studies should also be replicated in future studies with larger sample sizes, given the small sizes of each group (ranging from 12 to 33). Moreover, while seven studies included a longitudinal component, except for two short clinical trials, none of the longitudinal studies systematized antidepressant treatment, making it impossible to assess the effect of time vs. treatment. Lastly, our review focused on structural MRI, and did not address functional neuroimaging modalities (e.g., functional MRI or positron emission tomography imaging); therefore, we cannot comment on the relationships between functional brain activity and cognitive function. Notwithstanding these limitations, some of the broad themes emerging from the reviewed studies deserve further discussion.
The cingulate region is a potential treatment target for LLD and LLD + MCI
One theme consistent across both T1 and DWI studies is the regional involvement of the cingulate (anterior and posterior sub-regions) region in impaired cognition in LLD in general [46, 53, 60, 63] and LOD in particular [66, 67]. The ACC region belongs to the salience network and has a unique role in emotional regulation and cognitive function due to connections to limbic (e.g., amygdala, hippocampus, striatum) and prefrontal cortex (e.g., DLPFC) areas . It has been suggested that emotional disturbances, particularly apathy, and cognitive deficits, particularly executive dysfunction, in LLD could share neurobiological mechanisms including poor white matter integrity of the ACC . The PCC is a corticolimbic structure and a core region of the default mode network (DMN) with connections to key regions implicated in memory such as the hippocampus, parahippocampus, and entorhinal cortex . This finding is congruent with results of several functional connectivity studies, which have consistently reported altered functional connectivity within the DMN in patients with LLD [80, 81], MCI + LLD , or AD with depressive symptoms . Moreover, this agrees with previous findings from our group indicating that posterior DMN regions (PCC and precuneus) consistently showed structural and functional alterations in LLD, MCI, and LLD + MCI . In a recent systematic review of 14 deep brain stimulation (DBS) clinical trials in patients with treatment-resistant depression, targeting the subcallosal cingulate cortex appeared to have a promising antidepressant effect [83, 84]. Taken together, findings from this review focused on depression and findings from studies focusing on MCI  suggest that the cingulate region is a region with shared vulnerability to both depression and cognitive impairment. It could be an intervention target for brain stimulation to mitigate the risk of AD associated with LLD or MCI.
White matter integrity is primarily associated with executive function
WM tract integrity was more strongly associated with executive function or processing speed, than learning or memory. Eight of the 10 reviewed DWI ROI studies found an association between WM tract integrity and executive function or processing speed [62, 65,66,67,68,69,70, 73]. By contrast, of the DWI studies, 6/8 ROI [64, 66,67,68, 70, 71] and 2/4 TBSS studies [59, 61] did not find an association between WM integrity and memory or learning. This finding holds true across different studies and within studies that tested different cognitive domains alongside each other [59, 65, 70]. This finding in LLD is congruent with studies in healthy older adults [86,87,88]. This suggests that degradation of WM integrity associated with normal aging or LLD drives the observed higher-order cognitive decline in executive function and processing speed.
LOD is a potential marker for cognitive decline
Lastly, stronger brain-cognition relationships were observed in LOD than EOD or non-depressed controls. This result extends findings from our meta-analysis of structural MRI studies showing more abnormalities in the gray matter of fronto-parietal, dorsal attention, and visual networks in LOD than in those of EOD or mixed-onset LLD . Similarly, in a longitudinal study of hippocampal volumes, participants with LOD experienced hippocampal atrophy at a faster rate than those with EOD . Other studies have also reported a strong association between white matter hyperintensities (WMH) in fronto-striatal circuits and cognitive decline in LOD . These findings support the hypothesis that LOD is a form of “vascular depression” and a prodrome for dementia [17, 90, 91]. The differences in brain-cognition relationship between LOD and EOD in the studies we reviewed are consistent with this hypothesis.
Our review of 51 studies of brain-cognition relationships using T1 and DWI measures provides evidence of brain circuitry that could differentially underlie cognitive impairment in LLD. Our analysis of longitudinal T1 studies highlights the role of the hippocampus in global cognition and other domains of memory in LLD, while our cross-sectional analysis highlights the role of additional corticolimbic regions including the cingulate cortex in learning and memory. Findings from the DWI studies we reviewed implicate white matter integrity of the cingulate bundle sub-regions with executive dysfunction. Our results highlight gray matter regions and white matter circuitry with a shared vulnerability to both LLD and cognitive impairment and summarize structural brain circuitry vulnerable to cognitive impairment in LLD. Our results may inform the design of preventive interventions for patients at risk of developing AD.
While a reasonable number of studies have explored the relationship between GM or WM and cognitive impairment in LLD, many questions remain unaddressed in this field. Future studies should include and contrast participants with LLD and with varying degrees of cognitive impairment to determine whether brain-cognition relationships differ between LLD and LLD + MCI. In our review, only 4 studies compared brain-cognition relationships in EOD versus LOD; future studies should stratify their analysis according to depression age of onset. Also, most of the GM studies and all the WM studies we reviewed were cross-sectional; future longitudinal studies using multi-modal imaging that control for antidepressant treatment are needed to better understand causal relationships between alterations in brain circuitry and cognitive impairment. These future studies will need larger sample sizes to identify reproducible brain-behavior associations . By uncovering altered brain circuitry and cognition in specific subgroups of patients with LLD and varying degrees of cognitive impairment, these future studies will inform the design of strategies to prevent AD.
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.
Blazer DG. Depression in late life: review and commentary. J Gerontol A Biol Sci Med Sci. 2003;58:249–65.
Bhalla RK, Butters MA, Becker JT, Houck PR, Snitz BE, Lopez OL, et al. Patterns of mild cognitive impairment after treatment of depression in the elderly. Am J Geriatr Psychiatry. 2009;17:308–16.
Butters MA, Whyte EM, Nebes RD, Begley AE, Dew MA, Mulsant BH, et al. The nature and determinants of neuropsychological functioning in late-life depression. Arch Gen Psychiatry. 2004;61:587.
Nebes RD, Butters MA, Mulsant BH, Pollock BG, Zmuda MD, Houck PR, et al. Decreased working memory and processing speed mediate cognitive impairment in geriatric depression. Psychol Med. 2000;30:679–91.
Dybedal GS, Tanum L, Sundet K, Gaarden TL, Bjølseth TM. Neuropsychological functioning in late-life depression. Front Psychol. 2013. https://doi.org/10.3389/fpsyg.2013.00381.
Koenig AM, Bhalla RK, Butters MA. Cognitive functioning and late-life depression. J Int Neuropsychol Soc. 2014;20:461–7.
Sexton CE, McDermott L, Kalu UG, Herrmann LL, Bradley KM, Allan CL, et al. Exploring the pattern and neural correlates of neuropsychological impairment in late-life depression. Psychol Med. 2012;42:1195–202.
Sheline YI, Barch DM, Garcia K, Gersing K, Pieper C, Welsh-Bohmer K, et al. Cognitive function in late life depression: relationships to depression severity, cerebrovascular risk factors and processing speed. Biol Psychiatry. 2006;60:58–65.
Diniz BS, Butters MA, Albert SM, Dew MA, Reynolds CF. Late-life depression and risk of vascular dementia and Alzheimer’s disease: systematic review and meta-analysis of community-based cohort studies. Br J Psychiatry J Ment Sci. 2013;202:329–35.
Barnes DE, Yaffe K, Byers AL, McCormick M, Schaefer C, Whitmer RA. Midlife vs late-life depressive symptoms and risk of dementia: differential effects for Alzheimer disease and vascular dementia. Arch Gen Psychiatry. 2012;69:493–8.
Byers AL, Yaffe K. Depression and risk of developing dementia. Nat Rev Neurol. 2011;7:323–31.
Geerlings MI, den Heijer T, Koudstaal PJ, Hofman A, Breteler MMB. History of depression, depressive symptoms, and medial temporal lobe atrophy and the risk of Alzheimer disease. Neurology. 2008;70:1258–64.
Dillon C. Late- versus early-onset geriatric depression in a memory research center. Neuropsychiatr Dis Treat. 2009;5:517.
Sachs-Ericsson N, Corsentino E, Moxley J, Hames JL, Rushing NC, Sawyer K, et al. A longitudinal study of differences in late- and early-onset geriatric depression: depressive symptoms and psychosocial, cognitive, and neurological functioning. Aging Ment Health. 2013;17:1–11.
Herrmann LL, Le Masurier M, Ebmeier KP. White matter hyperintensities in late life depression: a systematic review. J Neurol Neurosurg Psychiatry. 2008;79:619–24.
Butters MA, Young JB, Lopez O, Aizenstein HJ, Mulsant BH, Reynolds CF, et al. Pathways linking late-life depression to persistent cognitive impairment and dementia. Dialogues Clin Neurosci. 2008;10:345–57.
Alexopoulos GS. Mechanisms and treatment of late-life depression. Transl Psychiatry. 2019;9:188.
Mahgoub N, Alexopoulos GS. Amyloid hypothesis: is there a role for antiamyloid treatment in late-life depression? Am J Geriatr Psychiatry J Am Assoc Geriatr Psychiatry. 2016;24:239–47.
Taylor WD, Aizenstein HJ, Alexopoulos GS. The vascular depression hypothesis: mechanisms linking vascular disease with depression. Mol Psychiatry. 2013;18:963–74.
Taylor WD, Zald DH, Felger JC, Christman S, Claassen DO, Horga G, et al. Influences of dopaminergic system dysfunction on late-life depression. Mol Psychiatry. 2022;27:180–91.
Li W, Muftuler LT, Chen G, Ward BD, Budde MD, Jones JL, et al. Effects of the coexistence of late-life depression and mild cognitive impairment on white matter microstructure. J Neurol Sci. 2014;338:46–56.
Rashidi-Ranjbar N, Miranda D, Butters MA, Mulsant BH, Voineskos AN. Evidence for structural and functional alterations of frontal-executive and corticolimbic circuits in late-life depression and relationship to mild cognitive impairment and dementia: a systematic review. Front Neurosci. 2020;14:253.
Xie C, Li W, Chen G, Douglas Ward B, Franczak MB, Jones JL, et al. The co-existence of geriatric depression and amnestic mild cognitive impairment detrimentally affect gray matter volumes: voxel-based morphometry study. Behav Brain Res. 2012;235:244–50.
Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.
Nuninga JO, Claessens TFI, Somers M, Mandl R, Nieuwdorp W, Boks MP, et al. Immediate and long-term effects of bilateral electroconvulsive therapy on cognitive functioning in patients with a depressive disorder. J Affect Disord. 2018;238:659–65.
Droppa K, Karim HT, Tudorascu DL, Karp JF, Reynolds CF, Aizenstein HJ, et al. Association between change in brain gray matter volume, cognition, and depression severity: Pre- and post- antidepressant pharmacotherapy for late-life depression. J Psychiatr Res. 2017;95:129–34.
Hou Z, Yuan Y, Zhang Z, Bai F, Hou G, You J. Longitudinal changes in hippocampal volumes and cognition in remitted geriatric depressive disorder. Behav Brain Res. 2012;227:30–35.
Köhler S, Thomas AJ, Lloyd A, Barber R, Almeida OP, O’Brien JT. White matter hyperintensities, cortisol levels, brain atrophy and continuing cognitive deficits in late-life depression. Br J Psychiatry. 2010;196:143–9.
Marano CM, Workman CI, Lyman CH, Munro CA, Kraut MA, Smith GS. Structural imaging in late-life depression: association with mood and cognitive responses to antidepressant treatment. Am J Geriatr Psychiatry. 2015;23:4–12.
Sachs-Ericsson N, Sawyer K, Corsentino E, Collins N, Steffens DC. The moderating effect of the APOE ɛ4 allele on the relationship between hippocampal volume and cognitive decline in older depressed patients. Am J Geriatr Psychiatry. 2011;19:23–32.
Sawyer K, Corsentino E, Sachs-Ericsson N, Steffens DC. Depression, hippocampal volume changes, and cognitive decline in a clinical sample of older depressed outpatients and non-depressed controls. Aging Ment Health. 2012;16:753–62.
Steffens DC, McQuoid DR, Payne ME, Potter GG. Change in hippocampal volume on magnetic resonance imaging and cognitive decline among older depressed and nondepressed subjects in the neurocognitive outcomes of depression in the elderly study. Am J Geriatr Psychiatry. 2011;19:4–12.
Ashtari M, Greenwald BS, Kramer-Ginsberg E, Hu J, Wu H, Patel M, et al. Hippocampal/amygdala volumes in geriatric depression. Psychol Med. 1999;29:629–38.
Greenwald BS, Kramer-Ginsberg E, Bogerts B, Ashtari M, Aupperle P, Wu H, et al. Qualitative magnetic resonance imaging findings in geriatric depression. Possible link between later-onset depression and Alzheimer’s disease? Psychol Med. 1997;27:421–31.
Lebedeva A, Borza T, Håberg AK, Idland A-V, Dalaker TO, Aarsland D, et al. Neuroanatomical correlates of late-life depression and associated cognitive changes. Neurobiol Aging. 2015;36:3090–9.
Steffens DC, Byrum CE, McQuoid DR, Greenberg DL, Payne ME, Blitchington TF, et al. Hippocampal volume in geriatric depression. Biol Psychiatry. 2000;48:301–9.
Almeida OP, Burton EJ, Ferrier N, McKEITH IG, O’Brien JT. Depression with late onset is associated with right frontal lobe atrophy. Psychol Med. 2003;33:675–81.
Chang C-C, Yu S-C, McQuoid DR, Messer DF, Taylor WD, Singh K, et al. Reduction of dorsolateral prefrontal cortex gray matter in late-life depression. Psychiatry Res Neuroimaging. 2011;193:1–6.
Bell-McGinty S, Butters MA, Meltzer CC, Greer PJ, Reynolds CF, Becker JT. Brain morphometric abnormalities in geriatric depression: long-term neurobiological effects of illness duration. Am J Psychiatry. 2002;159:1424–7.
Colloby SJ, Firbank MJ, Vasudev A, Parry SW, Thomas AJ, O’Brien JT. Cortical thickness and VBM-DARTEL in late-life depression. J Affect Disord. 2011;133:158–64.
Lai T-J, Payne ME, Byrum CE, Steffens DC, Krishnan KRR. Reduction of orbital frontal cortex volume in geriatric depression. Biol Psychiatry. 2000;48:971–5.
Lloyd AJ, Ferrier IN, Barber R, Gholkar A, Young AH, O’Brien JT. Hippocampal volume change in depression: late- and early-onset illness compared. Br J Psychiatry J Ment Sci. 2004;184:488–95.
Pantel J, Schröder J, Essig M, Popp D, Dech H, Knopp MV, et al. Quantitative magnetic resonance imaging in geriatric depression and primary degenerative dementia. J Affect Disord. 1997;42:69–83.
Lim HK, Jung WS, Ahn KJ, Won WY, Hahn C, Lee SY, et al. Regional cortical thickness and subcortical volume changes are associated with cognitive impairments in the drug-naive patients with late-onset depression. Neuropsychopharmacology. 2012;37:838–49.
Shin J-H, Um YH, Lee CU, Lim HK, Seong J-K. Multiple cortical thickness sub-networks and cognitive impairments in first episode, drug naïve patients with late life depression: a graph theory analysis. J Affect Disord. 2018;229:538–45.
Elderkin-Thompson V, Hellemann G, Pham D, Kumar A. Prefrontal brain morphology and executive function in healthy and depressed elderly. Int J Geriatr Psychiatry. 2009;24:459–68.
Sheline YI, Price JL, Vaishnavi SN, Mintun MA, Barch DM, Epstein AA, et al. Regional white matter hyperintensity burden in automated segmentation distinguishes late-life depressed subjects from comparison subjects matched for vascular risk factors. Am J Psychiatry. 2008;165:524–32.
Yuan Y, Zhu W, Zhang Z, Bai F, Yu H, Shi Y, et al. Regional gray matter changes are associated with cognitive deficits in remitted geriatric depression: an optimized voxel-based morphometry study. Biol Psychiatry. 2008;64:541–4.
Ballmaier M, Narr KL, Toga AW, Elderkin-Thompson V, Thompson PM, Hamilton L, et al. Hippocampal morphology and distinguishing late-onset from early-onset elderly depression. Am J Psychiatry. 2008;165:229–37.
Choi WH, Jung WS, Um YH, Lee CU, Park YH, Lim HK. Cerebral vascular burden on hippocampal subfields in first-onset drug-naïve subjects with late-onset depression. J Affect Disord. 2017;208:47–53.
Avila R, Ribeiz S, Duran FLS, Arrais JPJ, Moscoso MAA, Bezerra DM, et al. Effect of temporal lobe structure volume on memory in elderly depressed patients. Neurobiol Aging. 2011;32:1857–67.
Egger K, Schocke M, Weiss E, Auffinger S, Esterhammer R, Goebel G, et al. Pattern of brain atrophy in elderly patients with depression revealed by voxel-based morphometry. Psychiatry Res Neuroimaging. 2008;164:237–44.
Lamar M, Charlton R, Zhang A, Kumar A. Differential associations between types of verbal memory and prefrontal brain structure in healthy aging and late life depression. Neuropsychologia. 2012;50:1823–9.
Steffens DC, McQuoid DR, Welsh-Bohmer KA, Krishnan KRR. Left orbital frontal cortex volume and performance on the benton visual retention test in older depressives and controls. Neuropsychopharmacology. 2003;28:2179–83.
Jayaweera HK, Hickie IB, Duffy SL, Mowszowski L, Norrie L, Lagopoulos J, et al. Episodic memory in depression: the unique contribution of the anterior caudate and hippocampus. Psychol Med. 2016;46:2189–99.
Yuan Y, Zhang Z, Bai F, You J, Yu H, Shi Y, et al. Genetic variation in apolipoprotein E alters regional gray matter volumes in remitted late-onset depression. J Affect Disord. 2010;121:273–7.
Dahabra S, Ashton CH, Bahrainian M, Britton PG, Ferrier IN, McAllister VA, et al. Structural and functional abnormalities in elderly patients clinically recovered from early- and late-onset depression. Biol Psychiatry. 1998;44:34–46.
Mai N, Zhong X, Chen B, Peng Q, Wu Z, Zhang W, et al. Weight rich-club analysis in the white matter network of late-life depression with memory deficits. Front Aging Neurosci. 2017;9:279.
He X, Pueraro E, Kim Y, Garcia CM, Maas B, Choi J, et al. Association of white matter integrity with executive function and antidepressant treatment outcome in patients with late-life depression. Am J Geriatr Psychiatry. 2021;29:1188–98.
Murphy CF, Gunning-Dixon FM, Hoptman MJ, Lim KO, Ardekani B, Shields JK, et al. White-matter integrity predicts stroop performance in patients with geriatric depression. Biol Psychiatry. 2007;61:1007–10.
Yuan Y, Zhang Z, Bai F, Yu H, Shi Y, Qian Y, et al. White matter integrity of the whole brain is disrupted in first-episode remitted geriatric depression. NeuroReport. 2007;18:1845–9.
Lamar M, Charlton RA, Ajilore O, Zhang A, Yang S, Barrick TR, et al. Prefrontal vulnerabilities and whole brain connectivity in aging and depression. Neuropsychologia. 2013;51:1463–70.
Alves GS, Karakaya T, Fußer F, Kordulla M, O’Dwyer L, Christl J, et al. Association of microstructural white matter abnormalities with cognitive dysfunction in geriatric patients with major depression. Psychiatry Res Neuroimaging. 2012;203:194–200.
Charlton RA, Lamar M, Zhang A, Yang S, Ajilore O, Kumar A. White-matter tract integrity in late-life depression: associations with severity and cognition. Psychol Med. 2014;44:1427–37.
Mettenburg JM, Benzinger TL, Shimony JS, Snyder AZ, Sheline YI. Diminished performance on neuropsychological testing in late life depression is correlated with microstructural white matter abnormalities. NeuroImage. 2012;60:2182–90.
Yin Y, He X, Xu M, Hou Z, Song X, Sui Y, et al. Structural and functional connectivity of default mode network underlying the cognitive impairment in late-onset depression. Sci Rep. 2016;6:37617.
Yuan Y, Hou Z, Zhang Z, Bai F, Yu H, You J, et al. Abnormal integrity of long association fiber tracts is associated with cognitive deficits in patients with remitted geriatric depression: a cross-sectional, case-control study. J Clin Psychiatry. 2010;71:1386–90.
Shimony JS, Sheline YI, D’Angelo G, Epstein AA, Benzinger TLS, Mintun MA, et al. Diffuse microstructural abnormalities of normal-appearing white matter in late life depression: a diffusion tensor imaging study. Biol Psychiatry. 2009;66:245–52.
Alexopoulos GS, Kiosses DN, Choi SJ, Murphy CF, Lim KO. Frontal white matter microstructure and treatment response of late-life depression: a preliminary study. Am J Psychiatry. 2002;159:1929–32.
Zhou H, Zhong X, Chen B, Wang Q, Zhang M, Mai N, et al. Elevated homocysteine levels, white matter abnormalities and cognitive impairment in patients with late-life depression. Front Aging Neurosci. 2022;14:931560.
Wang Z, Yuan Y, You J, Zhang Z. Disrupted structural brain connectome underlying the cognitive deficits in remitted late-onset depression. Brain Imaging Behav. 2020;14:1600–11.
Li X, Steffens DC, Potter GG, Guo H, Song S, Wang L. Decreased between‐hemisphere connectivity strength and network efficiency in geriatric depression. Hum Brain Mapp. 2017;38:53–67.
Wang Z, Yuan Y, Jiang Y, You J, Zhang Z. Identification of specific neural circuit underlying the key cognitive deficit of remitted late-onset depression: A multi-modal MRI and machine learning study. Prog Neuropsychopharmacol Biol Psychiatry. 2021;108:110192.
Rajji TK, Miranda D, Mulsant BH, Lotz M, Houck P, Zmuda MD, et al. The MMSE is not an adequate screening cognitive instrument in studies of late-life depression. J Psychiatr Res. 2009;43:464–70.
Spencer RJ, Wendell CR, Giggey PP, Katzel LI, Lefkowitz DM, Siegel EL, et al. Psychometric limitations of the mini-mental state examination among nondemented older adults: an evaluation of neurocognitive and magnetic resonance imaging correlates. Exp Aging Res. 2013;39:382–97.
Marek S, Tervo-Clemmens B, Calabro FJ, Montez DF, Kay BP, Hatoum AS, et al. Reproducible brain-wide association studies require thousands of individuals. Nature. 2022;603:654–60.
Gratton C, Nelson SM, Gordon EM. Brain-behavior correlations: two paths toward reliability. Neuron. 2022;110:1446–9.
Stevens FL, Hurley RA, Taber KH. Anterior cingulate cortex: unique role in cognition and emotion. J Neuropsychiatry Clin Neurosci. 2011;23:121–5.
Leech R, Sharp DJ. The role of the posterior cingulate cortex in cognition and disease. Brain. 2014;137:12–32.
Guan C, Amdanee N, Liao W, Zhou C, Wu X, Zhang X et al. Altered intrinsic default mode network functional connectivity in patients with remitted geriatric depression and amnestic mild cognitive impairment. Int Psychogeriatr. 2021;34:703–714.
Jiang W-H, Yuan Y-G, Zhou H, Bai F, You J-Y, Zhang Z-J. Abnormally altered patterns of whole brain functional connectivity network of posterior cingulate cortex in remitted geriatric depression: a longitudinal study. CNS Neurosci Ther. 2014;20:772–7.
Zhang J, Guo Z, Liu X, Jia X, Li J, Li Y, et al. Abnormal functional connectivity of the posterior cingulate cortex is associated with depressive symptoms in patients with Alzheimer’s disease. Neuropsychiatr Dis Treat. 2017;13:2589–98.
Mayberg HS, Lozano AM, Voon V, McNeely HE, Seminowicz D, Hamani C, et al. Deep brain stimulation for treatment-resistant depression. Neuron. 2005;45:651–60.
Sobstyl M, Kupryjaniuk A, Prokopienko M, Rylski M. Subcallosal cingulate cortex deep brain stimulation for treatment-resistant depression: a systematic review. Front Neurol. 2022;13:780481.
Chandra A, Dervenoulas G, Politis M, Alzheimer’s Disease Neuroimaging Initiative. Magnetic resonance imaging in Alzheimer’s disease and mild cognitive impairment. J Neurol. 2019;266:1293–302.
Borghesani PR, Madhyastha TM, Aylward EH, Reiter MA, Swarny BR, Warner Schaie K, et al. The association between higher order abilities, processing speed, and age are variably mediated by white matter integrity during typical aging. Neuropsychologia. 2013;51:1435–44.
Haász J, Westlye ET, Fjær S, Espeseth T, Lundervold A, Lundervold AJ. General fluid-type intelligence is related to indices of white matter structure in middle-aged and old adults. NeuroImage. 2013;83:372–83.
Ystad M, Hodneland E, Adolfsdottir S, Haász J, Lundervold AJ, Eichele T, et al. Cortico-striatal connectivity and cognition in normal aging: a combined DTI and resting state fMRI study. NeuroImage. 2011;55:24–31.
Zhukovsky P, Anderson JAE, Coughlan G, Mulsant BH, Cipriani A, Voineskos AN. Coordinate-based network mapping of brain structure in major depressive disorder in younger and older adults: a systematic review and meta-analysis. Am J Psychiatry. 2021;178:1119–28.
Bennett S, Thomas AJ. Depression and dementia: cause, consequence or coincidence? Maturitas. 2014;79:184–90.
Sneed JR, Culang-Reinlieb ME. The vascular depression hypothesis: an update. Am J Geriatr Psychiatry. 2011;19:99–103.
Liu S, Abdellaoui A, Verweij KJH, Van Wingen GA. Replicable brain–phenotype associations require large-scale neuroimaging data. Nat Hum Behav. https://doi.org/10.1038/s41562-023-01642-5 (2023.).
The authors would like to acknowledge financial support from the Queen Elizabeth II/Gregory M. Brown Scholarship in Science and Technology and Labatt Family Chair in Biology of Depression in Late-Life Adults at the University of Toronto.
The authors declare no competing interests.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Marawi, T., Ainsworth, N.J., Zhukovsky, P. et al. Brain-cognition relationships in late-life depression: a systematic review of structural magnetic resonance imaging studies. Transl Psychiatry 13, 284 (2023). https://doi.org/10.1038/s41398-023-02584-2