Relationships Among Short Self-Reported Sleep Duration,Cognitive Impairment,and Insular Functional Connectivity in Late-Life Depression

Abstract

Background:

Both late-life depression (LLD) and short sleep duration increase the risk of cognitive impairment. Increased insular resting-state functional connectivity (FC) has been reported in individuals with short sleep duration and dementia.

Objective:

This study aimed to investigate whether short sleep duration is associated with impaired cognition and higher insular FC in patients with LLD.

Methods:

This case– control study recruited 186 patients with LLD and 83 normal controls (NC), and comprehensive psychometric assessments, sleep duration reports and resting-state functional MRI scans (81 LLD patients and 54 NC) were conducted.

Results:

Patients with LLD and short sleep duration (LLD-SS patients) exhibited more severe depressive symptoms and worse cognitive function than those with normal sleep duration (LLD-NS patients) and NC. LLD-SS patients exhibited higher FC between the bilateral insula and inferior frontal gyrus (IFG) pars triangularis than LLD-NS patients and NC, while LLD-NS patients exhibited lower FC than NC. Increased insular FC was correlated with short sleep duration, severe depressive symptoms, and slower information processing speeds. Furthermore, an additive effect was found between sleep duration and LLD on global cognition and insular FC.

Conclusion:

LLD-SS patients exhibited impaired cognition and increased insular FC. Abnormal FC in LLD-SS patients may be a therapeutic target for neuromodulation to improve sleep and cognitive performance and thus decrease the risk of dementia.

Keywords

Alzheimer’s disease cognitive impairment functional connectivity late-life depression MRI sleep duration

INTRODUCTION

Late-life depression (LLD) is defined as major depressive disorder in the geriatric population and affects 1.8% to 7.2% of geriatric adults [1, 2]. LLD is associated with impaired cognition [3] and a high risk of dementia [4]. Specifically, longitudinal studies have shown that the incidence of Alzheimer’s disease (AD) is increased in patients with LLD [5 –7]. LLD and AD share some pathophysiological mechanisms, including atrophy of both frontocortical and limbic networks [8 –11], glutamatergic excitotoxicity [12, 13], dysregulation of the HPA axis [14, 15], decreased levels of brain-derived neurotrophic factor [16, 17], and neuroinflammation [18]. Therefore, early identification (and intervention) regarding risk factors for dementia in LLD patients may prevent the development of dementia.

Nocturnal sleep is vital across the lifespan, and a sleep duration (SD) of 7 to 8 hours is recommended for older adults [19]. Poor sleep increases the risk of depression [20], cognitive decline [21, 22], and AD [23, 24]. Self-reported short SD (≤6 h/night) is significantly associated with worse cognitive performance [25, 26] and a higher burden of amyloid-β (a biomarker of AD) [27]. Additionally, short SD may precede cognitive decline in AD [28]. At any age, inadequate SD is one of the most common symptoms of depression [29]. Self-reported sleep disturbances were observed in 59.9% of LLD patients and associated with depressive symptoms [30]. Furthermore, previous studies have indicated that sleep disturbances, including insomnia, were associated with poor outcomes, poor response to depression treatment [31, 32], and a high recurrence risk [33]. However, whether short SD, as a modifiable risk factor, is associated with impaired cognition remains unclear in the context of LLD.

Although the pathophysiology of short SD remains unclear, accumulating evidence indicates that the insula and salience network (SN) may play a key role [34, 35]. The insula, a core hub of the salience network, evaluates whether a stimulus is emotionally or motivationally salient and links external information with internal motivation [36]. In structural MRI studies, insufficient sleep was associated with changes in gray matter volume and fractional anisotropy of the insula [37, 38] and decreased structural connectivity in a network with the insula as a key node [39]. In a resting-state functional MRI study, increased FC of the insula was observed in individuals with insomnia [35] and in patients suffering from depression with insomnia complaints [40]; this increase in FC was associated with cognitive deficits [41]. According to a systematic review and meta-analysis, AD is associated with hyperconnectivity of the SN, especially in the anterior insula [42]. Increased FC of the insular cortex and SN has been reported in cognitively normal individuals with elevated amyloid levels [43], in carriers of APOE ɛ4 (a risk gene of AD) [44] and in patients with mild cognitive impairment (MCI) [45]. Abnormal insular FC is associated with impaired cognition in MCI [45] and AD [46]. Regarding LLD, reduced insular cortical thickness and gray matter volume were correlated with more severe apathy and anxiety symptoms [47, 48], respectively. Moreover, lower FC between the insula component of the SN and the left executive control network in patients with LLD was reported compared with healthy controls [49]. However, the insular FC of LLD patients with short SD and the relationships among abnormal insular FC, short SD, and impaired cognitive function in LLD patients remain unclear.

Both LLD and self-reported short SD are modifiable risk factors for developing dementia. Increased insular FC has been reported in individuals with short SD and is associated with cognitive decline along the AD continuum. Therefore, it is necessary to clarify whether LLD patients with self-reported short SD exhibit impaired cognition and higher insular FC than those with adequate SD. The present study compared the cognitive performance and insular FC patterns between LLD patients with short or adequate SD and explored the relationships among insular FC, short SD, and cognitive function. The present study may provide new insights into the pathophysiological substrate by which short SD increases the risk of dementia in patients with LLD.

METHODS

Participants

A total of 186 patients with LLD were recruited from the Affiliated Brain Hospital of Guangzhou Medical University and 83 normal controls (NC) matched according to age and sex distributions were recruited from the community between June 2016 and October 2020. Informed consent was obtained from all participants before participation in the study. The study was approved by the Ethics Committee of the Affiliated Brain Hospital of Guangzhou Medical University (2015, 013). The inclusion criteria for patients with LLD were as follows: 1) age≥55 years, 2) at least one episode of depression after 55 years, and 3) a diagnosis of major depressive disorder according to the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV). The inclusion criteria for NC were as follows: 1) no history of psychiatric disorders and 2) normal cognition.

The diagnoses and assessments of all participants were confirmed by at least two trained psychiatrists. Participants were excluded if they met the following criteria: 1) history of other major psychiatric illnesses, such as bipolar disorder or schizophrenia; 2) physical diseases that can cause emotional abnormalities, such as anemia or hypothyroidism; and 3) neurological diseases, such as Parkinson’s disease or stroke. All participants who underwent MRI were right-handed and had no contraindication that precluded MRI scans. The question “During the past month, how many hours of sleep did you get each night?” was used to measure sleep duration. According to self-reported SD in the past month, LLD patients were divided into two subgroups: the LLD-SS group (LLD patients with short SD:≤6 h) and the LLD-NS group (LLD patients with normal SD: 7-8 h) [27]. Patients with other values of self-reported SD were not included in the study.

Psychometric assessments

The participants underwent a comprehensive battery of psychometric assessments by a trained psychologist and a psychiatrist in a specific room. The battery assessed 1) global cognitive performance with the Mini-Mental State Examination (MMSE), 2) activities of daily living with the Activities of Daily Living Scale (ADL), 3) memory function with the Auditory Verbal Learning Test (N5, the recognition item), 4) visuospatial skills with the Rey-Osterrieth Complex Figure test, 5) working memory with the Digit Span Test, 6) language function with the Verbal Fluency Test, 7) executive function with Stroop Color and Word Test C, and 8) information processing speed (IPS) with the Symbolic Digit Transformation Test (SDMT).

Depressive symptoms were assessed with the Geriatric Depression Scale (GDS), a battery that lacks items assessing sleep disturbance, and the adjusted 17-item Hamilton Depression Rating Scale (HAMD), in which the scores on sleep-related items (items 4, 5 and 6) were excluded to minimize any effects of sleep disturbance on the severity of depression (below, HAMD scores refer to adjusted HAMD scores) [40, 50].

MRI acquisition and preprocessing

The participants underwent MRI scans on Philips 3.0T MR systems. A single-shot gradient echo-planar imaging (EPI) pulse sequence in 8 min was used to obtain a sagittal resting-state fMRI dataset. The parameters of the EPI pulse sequence were as follows: TE = 30 ms, TR = 2000 ms, volume = 240, flip angle (FA) = 90 degrees, number of slices = 33, slice thickness = 4 mm, matrix size = 64×64, and field of view (FOV) = 220×220 mm. Statistical Parametric Mapping (SPM12) and Data Processing Assistant for Resting-State 5.1 (DPASF 5.1) were used for data preprocessing [51]. After the first 10 images were discarded, the remaining 230 volumes were corrected for slice timing and head motion. Head motion data were provided after realignment correction. Participants were excluded if they had > 2° of angular motion, 2 mm of displacement in any plane, or 0.2 mm of mean framewise displacement (FD) (no participants were excluded). The standard Montreal Neurological Institute (MNI) EPI template was applied for spatial normalization, and the images were resampled to 3×3×3 mm³ and smoothed using a 6-mm full-width at half-maximum (FWHM) Gaussian kernel. After removing the linear trend, signals of white matter, cerebrospinal fluid, and the Friston-24 parameters of head motion were regressed out as nuisance covariates [52]. To minimize the influence of head motion, the mean FD of each participant was regressed out in group-level analysis. A bandpass filter of 0.01–0.1 Hz was applied [53].

Seed-based FC analysis

Regions of interest (ROI) in the seed-based FC analysis were the left and right insula, anatomically derived from the automated anatomical labeling atlas [54, 55]. The time-series data in each ROI were extracted and averaged for reference as the seed point. FC maps of each individual were acquired from the Pearson correlation coefficient between the averaged time series of each ROI and the time series of each voxel of the whole brain. Fisher’s r-to-z transformation was used to improve the normality of FC maps. Group-level statistical analyses of FC were based on z-score maps.

Statistical analyses

Statistical analyses were carried out using The Statistical Package for the Social Sciences (IBM SPSS 25.0). Demographic variables were analyzed by univariate ANOVA with general linear model and the χ² test. Psychometric variables were analyzed by ANCOVA with general linear model, and sex, age, and years of education were included as covariates. The least significant difference (LSD) method was used to determine significant differences between pairwise comparisons. All statistical tests were two-tailed, and the significance level was defined as p < 0.05.

For FC analyses, significant differences among the three groups were tested using one-way ANCOVA on z-score maps for each ROI. The FC maps were corrected using a false discovery rate (FDR) correction at q < 0.05. The significantly different regions among the three groups in ANCOVA were extracted and averaged for post hoc analysis with LSD correction. Control variables included mean FD, sex, age, and years of education.

Partial correlation analyses were applied to detect the relationships among SD, psychometric variables, and FC, controlling sex, age, and years of education. Multiple linear regression analyses were used to further explore the factors that determine SD-related cognitive performance and insular FC. Dependent variables included neuropsychological scores and FC respectively that significantly correlated with SD in correlation analyses, independent variables included SD, LLD, an interaction term between SD and LLD, and other variables that correlated with dependent variables. Specifically, the regression model was first constructed with covariates including age, sex, year of education and FD (FD was included when FC was a dependent variable). Then a stepwise analysis method was used for independent variables, with p(entry) 0.05 and p(removal) 0.10. Each potential independent variables was tested as it was added to the model, retaining those deemed most statistically significant. This process is repeated until the results are optimal.

Validation analysis

Depressive symptoms (HAMD scores) and insomnia symptoms (sleep items of the HAMD) were controlled to validate the comparison of cognitive performance and insular FC among the three groups.

RESULTS

Demographic and psychometric characteristics

Demographic and psychometric data are listed in Table 1. There was no significant difference in age or sex distribution among the three groups. The LLD-SS group had significantly fewer years of education and lower scores on the ADL, MMSE, and SMDT as well as higher HAMD scores than the other two groups (p < 0.05, LSD correction). The NC group exhibited significantly lower GDS and HAMD scores and higher scores on the MMSE, Auditory-Verbal Learning Test, Rey– Osterrieth Complex Figure test, Digit Span Test, Verbal Fluency Test, and SDMT than the other two groups (p < 0.05, LSD correction). No significant difference was found in scores on the Stroop Color and Word Test C among the three groups. Functional MRI data were collected and analyzed from 56 LLD-SS patients, 25 LLD-NS patients and 54 NC. There was no significant difference in FD among the three groups (LLD-SS patients: 0.051±0.021, LLD-NS patients: 0.053±0.029, NC: 0.047±0.023, F = 0.794, p = 0.454).

Table 1

Demographic and clinical data of all participants

	LLD-SS	LLD-NS	NC	F /χ²	p	Post-hoc^#
	(n = 133)	(n = 53)	(n = 83)
Age	68.82 (6.94)	67.42 (6.18)	68.39 (6.37)	0.93	0.398	–
Male/Female	28/105	14/39	25/58	2.33	0.312	–
Education	7.94 (3.63)	11.80 (8.10)	11.20 (3.18)	18.50	< 0.001	A<B, C
SD	4.92 (0.98)	7.56 (0.50)	6.35 (1.05)	166.28	< 0.001	A<C<B
Onset/remission	27/106	9/44	–	0.27	0.605	–
Age of first onset	59.67 (11.52)	57.40 (110.4)	–	1.33	0.251
HAMD^a	8.23 (6.34)	4.98 (4.30)	0.88(1.43)	51.69^*	< 0.001	A>B>C
GDS	5.57 (3.89)	4.15 (3.83)	1.71 (1.92)	19.00^*	< 0.001	A, B > C
ADL	12.74 (2.47)	13.62 (0.86)	13.98 (0.15)	10.72^*	< 0.001	A<B, C
MMSE	20.68 (4.88)	25.04 (3.00)	26.98 (1.63)	54.85^*	< 0.001	A<B<C
SDMT	22.08 (12.49)	27.98 (11.37)	34.96 (9.71)	36.57^*	< 0.001	A<B<C
The Auditory-Verbal Learning Test	3.76 (3.01)	4.75 (3.11)	6.90 (2.56)	27.15^*	< 0.001	A, B < C
Rey– Osterrieth Complex Figure	21.72 (7.64)	24.00 (6.27)	27.75 (3.32)	11.73^*	< 0.001	A, B < C
Digit Span Test	14.31 (3.72)	14.79 (3.41)	17.42 (3.39)	10.53^*	< 0.001	A, B < C
Verbal Fluency Test	11.93 (4.00)	12.85 (3.78)	15.31 (4.22)	16.65^*	< 0.001	A, B < C
Stroop Color and Word Test C	93.25 (29.67)	91.43 (32.18)	81.18 (22.78)	2.68^*	0.072	–

SD, self-report sleep duration; LLD-SS, late-life depression with short SD; LLD-NS, late-life depression with normal SD; NC, normal control; HAMD^a, adjusted Hamilton Depression Scale (sleep items were removed); GDS, Geriatric Depression Scale; ADL, activities of daily living; MMSE, Mini-Mental State Examination; DST, Digit Span Test; SDMT, Symbolic Digit Transformation Test. ^*Age, sex, and years of education were included as control variables. ^#A, LLD-SS; B, LLD-NS; C, NC.

Differences in insular FC among the LLD-SS, LLD-NS, and NC groups

The ANCOVA on insular FC revealed differences in FC among the LLD-SS, LLD-NS and NC groups (shown in Table 2 and Fig. 1A, B), including the FC (1) between the left insula and right inferior frontal gyrus (IFG) pars triangularis, and (2) between the right insula and right IFG pars triangularis. In the post hoc analysis (Fig. 1C), the LLD-SS group exhibited increased FC between the left insula and right IFG pars triangularis compared with the LLD-NS (p < 0.001) and NC group (p < 0.001) and increased FC between the right insula and right IFG pars triangularis compared with the LLD-NS (p < 0.001) and NC group (p < 0.001). The LLD-NS group exhibited decreased FC between the left insula and right IFG pars triangularis compared with NC group (p = 0.047) and between the right insula and right IFG pars triangularis compared with the HC group (p = 0.011).

Table 2

Differences of FC among LLD-SS, LLD-NS, and NC group

ROI	Regions	Clusters	Peak MNI			Peak F
		size(mm³)	Coordinates
			X	Y	Z
Left Insula	Right IFG pars triangularis	891	54	21	3	20.27
Right Insula	Right IFG pars triangularis	2673	54	24	6	27.34

False discovery rate (FDR) at q < 0.05. FC, functional connectivity; ROI, regions of interest. LLD-SS, late-life depression with short sleep duration; LLD-NS, late-life depression with normal sleep duration; NC, normal control; MNI, Montreal Neurological Institute; IFG, inferior frontal gyrus.

Fig. 1

Differences of FC among LLD-SS, LLD-NS, and NC group. A) Differences of FC between left insula and right IFG pars triangularis; B) Differences of FC between right insula and right IFG pars triangularis; C) Comparison of FC among LLD-SS, LLD-NS, and NC group: LLD-SS group exhibited increased FC between left insula and right IFG pars triangularis, and between right insula and right IFG pars triangularis compared with LLD-NS (p < 0.001, p < 0.001) and NC group (p < 0.001, p < 0.001). LLD-NS group exhibited decreased FC between left insula and right IFG pars triangularis, and between right insula and right IFG pars triangularis compared with NC group (p = 0.047, p = 0.011). ^*p < 0.05; ^**p < 0.01; ^***p < 0.001. FC, functional connectivity; LLD-SS, late-life depression with short sleep duration; LLD-NS, late-life depression with normal sleep duration; NC, normal control; IFG, inferior frontal gyrus.

Correlation and regression analyses

In all participants, SD was correlated with HAMD (r = –0.26, p < 0.001), GDS (r = –0.21, p = 0.006), ADL (r = 0.20, p = 0.001), MMSE (r = 0.39, p < 0.001) and SDMT scores (r = 0.28, p < 0.001). No significant correlation was found between SD and other psychometric characteristics. Furthermore, regression analyses of the psychometric variables exhibited that 1) MMSE was associated with SD (β= 0.23, p < 0.001), LLD (β=–0.29, p < 0.001), interaction between SD and LLD (β= 0.09, p = 0.049), and HAMD (β=–0.25, p < 0.001); and that 2) SDMT was associated with SD (β= 0.14, p = 0.007), LLD (β=–0.26, p < 0.001), and HAMD (β=–0.22, p < 0.001) (Table 3).

Table 3

Regression analyses of the psychometric variables and insular FC

	R²	Variables	β	t	p
MMSE	0.48	SD	0.23	4.60	< 0.001
		LLD^#	–0.29	–5.33	< 0.001
		SD^* LLD	0.09	1.98	0.049
		HAMD	–0.25	–4.73	< 0.001
SDMT	0.43	SD	0.14	2.72	0.007
		LLD	–0.26	–4.81	< 0.001
		HAMD	–0.22	–4.00	< 0.001
Left insular FC	0.22	SD	–0.20	–2.37	0.019
		LLD	0.22	2.59	0.011
		SD^* LLD	–0.21	–2.57	0.011
Right insular FC	0.32	SD	–0.28	–3.60	< 0.001
		LLD	0.20	2.38	0.019
		SD^* LLD	–0.22	–2.80	0.006
		SDMT	–0.19	2.14	0.034

^#a categorical variable representing whether the participants suffered LLD or not. ^*an interaction term between SD and LLD. LLD, late-life depression; FC, functional connectivity; SD, sleep duration; MMSE, Mini-Mental State Examination; SDMT, Symbol-Digit Modality Test.

In participants who underwent functional MRI scans, SD was correlated with FC between the left insula and right IFG pars triangularis (r = –0.27, p = 0.001) and FC between the right insula and right IFG pars triangularis (r = –0.38 p < 0.001); FC between the right insula and right IFG pars triangularis correlated with HAMD (r = 0.19, p = 0.030) and SDMT scores (r = –0.27, p = 0.002). Furthermore, regression analyses of the insular FC exhibited that 1) FC between the left insula and right IFG pars triangularis was associated with SD (β=–0.20, p = 0.019), LLD (β= 0.22, p = 0.011), interaction between SD and LLD (β=–0.21, p = 0.011), and that 2) FC between the right insula and right IFG pars triangularis was associated with SD (β=–0.28, p < 0.001), LLD (β= 0.20, p = 0.019), interaction between SD and LLD (β=–0.22, p = 0.006) and SDMT (β=–0.19, p = 0.034) (Table 3).

Validation results

The main results validated the comparison of cognitive performance and insular FC among the three groups (Supplementary Tables 1 and 2).

DISCUSSION

The present study investigated the insular FC of LLD patients with short SD and explored the influence of insular FC on the relationship between SD and cognition. The main results are as follows. First, LLD-SS patients exhibited more severe depressive symptoms and worse cognitive function (especially in global cognition, ADL, and IPS) than LLD-NS patients and NC. SD significantly correlated with global cognition and IPS after adjusting potential confounders. Second, LLD-SS patients exhibited higher FC between the bilateral insula and IFG pars triangularis than the LLD-NS and NC group, while LLD-NS patients showed lower FC between the bilateral insula and IFG pars triangularis than the NC group. And this increase in insular FC was correlated with shorter SD, more severe depressive symptoms and slower IPS. Third, an additive effect was found between SD and LLD on global cognition and insular FC.

Accumulating evidence has indicated a crucial role of amyloid-β clearance in the relationship between short SD and the risk of AD, as short SD reduces clearance [56, 57]. Consistent with these experimental findings, an epidemiological study found an association between self-reported short SD and a greater amyloid-β burden in community-dwelling older adults [58]. Another study with a sample size of 4,000 individuals also found that self-reported short SD correlated with a high amyloid-β burden, which was associated with cognitive decline [27]. It is well acknowledged that short SD and cognitive impairment are common manifestations of LLD and may be influenced by the severity of depression [29, 59], which may lead to an underestimation of the effects of short SD on impaired cognition caused by potential AD-related pathology. In the present study, LLD patients with self-reported short SD exhibited impaired cognition relative to LLD patients with adequate SD and NC, and SD correlated with cognitive performance after adjustment for potential confounders. Furthermore, an additive effect was found between SD and LLD on global cognition. The results suggest that LLD patients with self-reported short SD may be more vulnerable to cognitive decline caused by AD-related pathology. Therefore, intervention for short SD in patients with LLD should be emphasized. Furthermore, longitudinal studies are necessary to verify the risk of AD in LLD patients with short SD and whether interventions for short SD decrease the risk of AD.

We found that LLD patients with short SD exhibited increased FC between the bilateral insula and right IFG pars triangularis compared with the LLD-NS and NC group. In addition, these increased FC values were associated with short SD, severe depressive symptoms and poor cognitive performance. Similarly, increased activation of the insula and IFG were detected by assessing the amplitude of low-frequency fluctuations (ALFF) in major depressive disorder patients with insomnia complaints [40]. Studies have shown that the IFG is correlated with deliberative processing, response inhibition, emotion regulation and memory maintenance [60 –63]. In an fMRI study using a modular graph theoretic analysis, the largest homotopic module (defined as the insular module, containing the IFG pars triangularis) was disrupted in AD, and the corresponding gray matter concentration was decreased in AD patients compared with NC [64]. NC carrying APOE ɛ4 showed higher fractional ALFF in the insula and IFG compared with those not carrying APOE ɛ4, indicating that greater activation may act as a compensatory mechanism against the neural challenges arising from AD pathology [65]. Additionally, according to a meta-analysis, the anterior insula and IFG were the only regions with overlapping activation for both physiological and psychosocial stressors, and this activation was considered to be a global neural stress reaction [66]. And there was found an additive effect between SD and LLD on FC between the bilateral insula and right IFG pars triangularis in the present study. Given the above findings and, it could be inferred that the increased FC between the insula and IFG in the present study may be a neural stress reaction due to short SD, more severe depressive symptoms and potential AD-related pathology. Similar to previous studies [47 –49], the LLD-NS group showed lower FC between the bilateral insula and IFG pars triangularis than the NC group, but only the difference in FC between the left insula and right IFG pars triangularis was significantly lower than that in the NC group (p = 0.047) and therefore should be interpreted cautiously. The decreased FC may be a manifestation of lower insular function in LLD. In the present study, the sign of changes in insular FC was associated with SD, and the underlying neural mechanism should be further explored in future studies.

In the present study, IPS seemed to be the most vulnerable to LLD-induced changes, as IPS scores were significantly lower in LLD-SS patients than in LLD-NS patients and NC, and IPS was correlated with SD and insular FC after adjusting for potential confounders; other cognitive domains did not show the same pattern. Among various cognitive impairments in patients with LLD, slower IPS is considered as the most core domain, and it plays a critical role in downstream processes of other cognitive domains [67, 68]. Moreover, slower processing speed impairs the development of expectancies in antidepressant trials with LLD patients, which may help explain the lower antidepressant response among older adults [69]. Therefore, greater attention should be given to IPS in LLD-SS patients, and early intervention may improve their prognosis. Furthermore, a previous study reported that the activation of the IFG and anterior insula is often discounted as due to motor inhibition or orienting responses, rather than changes in speed– accuracy trade-offs [70], which may affect IPS. Therefore, the FC between the right insula and right IFG pars triangularis may be a therapeutic target for neuromodulation to alleviate cognitive and sleep problems in LLD patients.

There are a few limitations of our study. First, self-reported SD is a subjective measure used as a proxy for poor sleep. The use of this proxy is supported by previous studies that reported a correlation between high amyloid-β burden and indices of poor sleep quality, including subjective sleep quality [71], objective sleep efficiency and fragmentation [72, 73], and nonrapid eye movement slow wave activity [74]. In the present study, the statistical significance of associations decreased when insomnia was controlled as a covariate, which means insomnia may mediate the impacts of short SD on cognition and FC. Therefore, future studies with objective sleep measures should be performed to resolve this discrepancy. Second, variable types and doses of antidepressants may be confounding factors in the relationships among FC, cognitive function, and depressive symptoms. Third, the insula is a heterogeneous brain region that comprises three cytoarchitecturally distinct subregions [75]. Further studies using insula subregions (rather than the entire insula) as ROIs may provide more detailed insights into the role of insular FC in LLD-SS patients. Additionally, other brain areas may be correlated with LLD and short SD, which merit further exploration. Fourth, the results from the present study were based on cross-sectional data, and the causality of associations among short SD, insular FC and risk of dementia could not be established. Well-designed longitudinal studies are needed in the future.

In conclusion, the LLD-SS group exhibited worse global cognition, slower IPS, and increased FC between the bilateral insula and IFG pars triangularis compared with the LLD-NS and NC group, suggesting that LLD patients with short SD may have a higher risk of developing dementia. Moreover, this increased insular FC was associated with short SD and poor cognitive performance. In particular, an additive effect was found between sleep duration and LLD on global cognition and insular FC. Abnormal FC in LLD patients with short SD may be therapeutic targets for neuromodulation to improve sleep and cognitive performance and thus decrease the risk of dementia. Future studies are needed to investigate whether interventions to normalize insular FC in LLD patients with short SD could decrease the risk of developing dementia and clarify its underlying mechanisms.

Footnotes

ACKNOWLEDGMENTS

The authors have no acknowledgments to report.

FUNDING

This study was supported by a grant from the National Natural Science Foundation of China (No. 82171533, No. 82101508), Medical Scientific Technology Research Foundation of Guangdong Province of China (NO.A2020446), the Key Laboratory for Innovation Platform Plan, the Science and Technology Program of Guangzhou, China, the Science and Technology Plan Project of Guangdong Province (No. 2019B030316001), Key Medical Specialty Construction Project of Traditional Chinese Medical Science in the 13th Five-Year Plan of Guangdong Province; Key Medical Specialty Construction Project of Traditional Chinese Medical Science of Guangzhou (2020–2022). Guangzhou Municipal Key Discipline in Medicine (2021-2023), Guangzhou High-level Clinical Key Specialty, Guangzhou Research-oriented Hospital. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

The supplementary material is available in the electronic version of this article: .

References

Beekman

, Copeland

, Prince

(1999) Review of community prevalence of depression in later life. Br J Psychiatry 174, 307–311.

Luppa

, Sikorski

, Luck

, Ehreke

, Konnopka

, Wiese

, Weyerer

, König

, Riedel-Heller

(2012) Age- and gender-specific prevalence of depression in latest-life–systematic review and meta-analysis. J Affect Disord 136, 212–221.

Alexopoulos

(2019) Mechanisms and treatment of late-life depression. Transl Psychiatry 9, 188.

Fiske

, Wetherell

, Gatz

(2009) Depression in older adults. Annu Rev Clin Psychol 5, 363–389.

Andersen

, Lolk

, Kragh-Sorensen

, Petersen

, Green

(2005) Depression and the risk of Alzheimer disease. Epidemiology 16, 233–238.

Richard

, Reitz

, Honig

, Schupf

, Tang

, Manly

, Mayeux

, Devanand

, Luchsinger

(2013) Late-life depression, mild cognitive impairment, and dementia. JAMA Neurol 70, 374–382.

Singh-Manoux

, Dugravot

, Fournier

, Abell

, Ebmeier

, Kivimäki

, Sabia

(2017) Trajectories of depressive symptoms before diagnosis of dementia: A 28-year follow-up study. JAMA Psychiatry 74, 712–718.

Sacuiu

, Insel

, Mueller

, Tosun

, Mattsson

, Jack

Jr. , DeCarli

, Petersen

, Aisen

, Weiner

, Mackin

(2016) Chronic depressive symptomatology in mild cognitive impairment is associated with frontal atrophy rate which hastens conversion to Alzheimer dementia. Am J Geriatr Psychiatry 24, 126–135.

Sexton

, Allan

, Le Masurier

, McDermott

, Kalu

, Herrmann

, Mäurer

, Bradley

, Mackay

, Ebmeier

(2012) Magnetic resonance imaging in late-life depression: Multimodal examination of network disruption. Arch Gen Psychiatry 69, 680–689.

10.

Laakso

, Soininen

, Partanen

, Helkala

, Hartikainen

, Vainio

, Hallikainen

, Hänninen

, Riekkinen

Sr. , (1995) Volumes of hippocampus, amygdala and frontal lobes in the MRI-based diagnosis of early Alzheimer’s disease: Correlation with memory functions. J Neural Transm Park Dis Dement Sect 9, 73–86.

11.

Wang

, Zang

, He

, Liang

, Zhang

, Tian

, Wu

, Jiang

, Li

(2006) Changes in hippocampal connectivity in the early stages of Alzheimer’s disease: Evidence from resting state fMRI. Neuroimage 31, 496–504.

12.

Chi

, Yu

, Tan

(2014) Depression in Alzheimer’s disease: Epidemiology, mechanisms, and management. J Alzheimers Dis 42, 739–755.

13.

Tanaka

, Meguro

, Yamaguchi

, Ishii

, Watanuki

, Funaki

, Yamaguchi

, Yamadori

, Iwata

, Itoh

(2003) Decreased striatal D2 receptor density associated with severe behavioral abnormality in Alzheimer’s disease. Ann Nucl Med 17, 567–573.

14.

Duman

(2014) Neurobiology of stress, depression, and rapid acting antidepressants: Remodeling synaptic connections. Depress Anxiety 31, 291–296.

15.

Galts

CPC

, Bettio

LEB

, Jewett

, Yang

, Brocardo

, Rodrigues

ALS

, Thacker

, Gil-Mohapel

(2019) Depression in neurodegenerative diseases: Common mechanisms and current treatment options. Neurosci Biobehav Rev 102, 56–84.

16.

Duman

, Monteggia

(2006) A neurotrophic model for stress-related mood disorders. Biol Psychiatry 59, 1116–1127.

17.

Connor

, Young

, Yan

, Faull

, Synek

, Dragunow

(1997) Brain-derived neurotrophic factor is reduced in Alzheimer’s disease. Brain Res Mol Brain Res 49, 71–81.

18.

Santos

, Beckman

, Ferreira

(2016) Microglial dysfunction connects depression and Alzheimer’s disease. Brain Behav Immun 55, 151–165.

19.

Hirshkowitz

, Whiton

, Albert

, Alessi

, Bruni

, DonCarlos

, Hazen

, Herman

, Adams Hillard

, Katz

, Kheirandish-Gozal

, Neubauer

, O’Donnell

, Ohayon

, Peever

, Rawding

, Sachdeva

, Setters

, Vitiello

, Ware

(2015) National Sleep Foundation’s updated sleep duration recommendations: Final report. Sleep Health 1, 233–243.

20.

Franzen

, Buysse

(2008) Sleep disturbances and depression: Risk relationships for subsequent depression and therapeutic implications. Dialogues Clin Neurosci 10, 473–481.

21.

Devore

, Grodstein

, Duffy

, Stampfer

, Czeisler

, Schernhammer

(2014) Sleep duration in midlife and later life in relation to cognition. J Am Geriatr Soc 62, 1073–1081.

22.

Virta

, Heikkilä

, Perola

, Koskenvuo

, Räihä

, Rinne

, Kaprio

(2013) Midlife sleep characteristics associated with late life cognitive function. Sleep 36, 1533–1541, 1541a.

23.

Lim

, Yu

, Kowgier

, Schneider

, Buchman

, Bennett

(2013) Modification of the relationship of the apolipoprotein E ɛ4 allele to the risk of Alzheimer disease and neurofibrillary tangle density by sleep. JAMA Neurol 70, 1544–1551.

24.

Lysen

, Luik

, Ikram

, Tiemeier

, Ikram

(2020) Actigraphy-estimated sleep and 24-hour activity rhythms and the risk of dementia. Alzheimers Dement 16, 1259–1267.

25.

, Groeger

, Cheng

, Dijk

, Chee

(2016) Self-reported sleep duration and cognitive performance in older adults: A systematic review and meta-analysis. Sleep Med 17, 87–98.

26.

, Liang

, Zheng

, Shi

, Zhong

, Xie

(2020) Association between sleep duration and cognitive decline. JAMA Netw Open 3, e2013573.

27.

Winer

, Deters

, Kennedy

, Jin

, Goldstein-Piekarski

, Poston

, Mormino

(2021) Association of short and long sleep duration with amyloid-β burden and cognition in aging. JAMA Neurol 78, 1187–1196.

28.

, Lucey

, Holtzman

(2014) Sleep and Alzheimer disease pathology–a bidirectional relationship. Nat Rev Neurol 10, 115–119.

29.

Ismail

, Fischer

, McCall

(2013) What characterizes late-lifedepression? Psychiatr Clin North Am 36, 483–496.

30.

Peters van Neijenhof

RJG

, van Duijn

, Comijs

, van den Berg

, de Waal

MWM

, Oude Voshaar

, van der Mast

(2018) Correlates of sleep disturbances in depressed older persons: The Netherlands study of depression in older persons (NESDO). Aging Ment Health 22, 233–238.

31.

Perlis

, Smith

, Lyness

, Matteson

, Pigeon

, Jungquist

, Tu

(2006) Insomnia as a risk factor for onset of depression in the elderly. Behav Sleep Med 4, 104–113.

32.

Dombrovski

, Mulsant

, Houck

, Mazumdar

, Lenze

, Andreescu

, Cyranowski

, Reynolds

, 3rd (2007) Residual symptoms and recurrence during maintenance treatment of late-life depression. J Affect Disord 103, 77–82.

33.

Lee

, Cho

, Olmstead

, Levin

, Oxman

, Irwin

(2013) Persistent sleep disturbance: A risk factor for recurrent depression in community-dwelling older adults. Sleep 36, 1685–1691.

34.

Chen

, Chang

, Glover

, Gotlib

(2014) Increased insula coactivation with salience networks in insomnia. Biol Psychol 97, 1–8.

35.

Van

Someren EJW

(2021) Brain mechanisms of insomnia: New perspectives on causes and consequences. Physiol Rev 101, 995–1046.

36.

Cauda

, D’Agata

, Sacco

, Duca

, Geminiani

, Vercelli

(2011) Functional connectivity of the insula in the resting brain. Neuroimage 55, 8–23.

37.

Dai

, Jiang

, Zhang

, Nie

, Liu

, Pei

, Gong

, Hu

, Lu

, Zhan

(2018) Plasticity and susceptibility of brain morphometry alterations to insufficient sleep. Front Psychiatry 9, 266.

38.

Khalsa

, Hale

, Goldstone

, Wilson

, Mayhew

, Bagary

, Bagshaw

(2017) Habitual sleep durations and subjective sleep quality predict white matter differences in the human brain. Neurobiol Sleep Circadian Rhythms 3, 17–25.

39.

Jespersen

, Stevner

, Fernandes

, Sørensen

, Van Someren

, Kringelbach

, Vuust

(2020) Reduced structural connectivity in insomnia disorder. J Sleep Res 29, e12901.

40.

Liu

, Guo

, Lu

, Tang

, Fan

, Wang

, Liu

(2018) Increased salience network activity in patients with insomnia complaints in major depressive disorder. Front Psychiatry 9, 93.

41.

Schiel

, Holub

, Petri

, Leerssen

, Tamm

, Tahmasian

, Riemann

, Spiegelhalder

(2020) Affect and arousal in insomnia: Through a lens of neuroimaging studies. Curr Psychiatry Rep 22, 44.

42.

Badhwar

, Tam

, Dansereau

, Orban

, Hoffstaedter

, Bellec

(2017) Resting-state network dysfunction in Alzheimer’s disease: A systematic review and meta-analysis. Alzheimers Dement (Amst) 8, 73–85.

43.

Fredericks

, Sturm

, Brown

, Hua

, Bilgel

, Wong

, Resnick

, Seeley

(2018) Early affective changes and increased connectivity in preclinical Alzheimer’s disease. Alzheimers Dement (Amst) 10, 471–479.

44.

Goveas

, Xie

, Chen

, Li

, Ward

, Franczak

, Jones

, Antuono

, Li

(2013) Functional network endophenotypes unravel the effects of apolipoprotein E epsilon 4 in middle-aged adults. PLoS One 8, e55902.

45.

Chand

, Wu

, Hajjar

, Qiu

(2017) Interactions of the salience network and its subsystems with the default-mode and the central-executive networks in normal aging and mild cognitive impairment. Brain Connect 7, 401–412.

46.

Zhang

, Guan

, Zhang

, Sun

, Li

, Hu

, Li

, Ye

, Meng

, Huang

, Lin

, Wang

, Liu

, Li

(2022) Disrupted coupling between salience network segregation and glucose metabolism is associated with cognitive decline in Alzheimer’s disease - A simultaneous resting-state FDG-PET/fMRI study. Neuroimage Clin 34, 102977.

47.

Pimontel

, Solomonov

, Oberlin

, Kanellopoulos

, Bress

, Hoptman

, Alexopoulos

, Gunning

(2021) Cortical thickness of the salience network and change in apathy following antidepressant treatment for late-life depression. Am J Geriatr Psychiatry 29, 241–248.

48.

Laird

, Siddarth

, Krause-Sorio

, Kilpatrick

, Milillo

, Aguilar

, Narr

, Lavretsky

(2019) Anxiety symptoms are associated with smaller insular and orbitofrontal cortex volumes in late-life depression. J Affect Disord 256, 282–287.

49.

, Wang

, Ward

, Antuono

, Li

, Goveas

(2017) Intrinsic inter-network brain dysfunction correlates with symptom dimensions in late-life depression. J Psychiatr Res 87, 71–80.

50.

Troxel

, Kupfer

, Reynolds

3rd , Frank

, Thase

, Miewald

, Buysse

(2012) Insomnia and objectively measured sleep disturbances predict treatment outcome in depressed patients treated with psychotherapy or psychotherapy-pharmacotherapy combinations. J Clin Psychiatry 73, 478–485.

51.

Yan

, Wang

, Zuo

, Zang

(2016) DPABI: Data processing & analysis for (resting-state) brain imaging. Neuroinformatics 14, 339–351.

52.

Friston

, Williams

, Howard

, Frackowiak

, Turner

(1996) Movement-related effects in fMRI time-series. Magn Reson Med 35, 346–355.

53.

Soares

, Magalhaes

, Moreira

, Sousa

, Ganz

, Sampaio

, Alves

, Marques

, Sousa

(2016) A hitchhiker’s guide to functional magnetic resonance imaging. Front Neurosci 10, 515.

54.

Tzourio-Mazoyer

, Landeau

, Papathanassiou

, Crivello

, Etard

, Delcroix

, Mazoyer

, Joliot

(2002) Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage 15, 273–289.

55.

, Li

, Zhang

, Pan

, Gao

, Zhan

, Liu

, Shao

, Zhang

(2021) Altered insula-prefrontal functional connectivity correlates to decreased vigilant attention after total sleep deprivation. Sleep Med 84, 187–194.

56.

Ooms

, Overeem

, Besse

, Rikkert

, Verbeek

, Claassen

(2014) Effect of 1 night of total sleep deprivation on cerebrospinal fluid β-amyloid 42 in healthy middle-aged men: A randomized clinical trial. JAMA Neurol 71, 971–977.

57.

Kang

, Lim

, Bateman

, Lee

, Smyth

, Cirrito

, Fujiki

, Nishino

, Holtzman

(2009) Amyloid-beta dynamics are regulated by orexin and the sleep-wake cycle. Science 326, 1005–1007.

58.

Spira

, Gamaldo

, An

, Wu

, Simonsick

, Bilgel

, Zhou

, Wong

, Ferrucci

, Resnick

(2013) Self-reported sleep and β-amyloid deposition in community-dwelling older adults. JAMA Neurol 70, 1537–1543.

59.

Otte

, Gold

, Penninx

, Pariante

, Etkin

, Fava

, Mohr

, Schatzberg

(2016) Major depressive disorder. Nat Rev Dis Primers 2, 16065.

60.

Morawetz

, Bode

, Baudewig

, Kirilina

, Heekeren

(2016) Changes in effective connectivity between dorsal and ventral prefrontal regions moderate emotion regulation. Cereb Cortex 26, 1923–1937.

61.

Aron

, Robbins

, Poldrack

(2014) Inhibition and the right inferior frontal cortex: One decade on. Trends Cogn Sci 18, 177–185.

62.

Heatherton

, Wagner

(2011) Cognitive neuroscience of self-regulation failure. Trends Cogn Sci 15, 132–139.

63.

McLaren

, Sreenivasan

, Diamond

, Mitchell

, Van Dijk

, Deluca

, O’Brien

, Rentz

, Sperling

, Atri

(2012) Tracking cognitive change over 24 weeks with longitudinal functional magnetic resonance imaging in Alzheimer’s disease. Neurodegener Dis 9, 176–186.

64.

Chen

, Zhang

, Xie

, Chen

, Zhang

, Teng

, Li

(2013) Modular reorganization of brain resting state networks and its independent validation in Alzheimer’s disease patients. Front Hum Neurosci 7, 456.

65.

Lin

, Ren

, Lo

, Chapman

, Jacobs

, Baran

, Porsteinsson

, Foxe

(2017) Insula and inferior frontal gyrus’ activities protect memory performance against Alzheimer’s disease pathology in old age. J Alzheimers Dis 55, 669–678.

66.

Kogler

, Müller

, Chang

, Eickhoff

, Fox

, Gur

, Derntl

(2015) Psychosocial versus physiological stress - Meta-analyses on deactivations and activations of the neural correlates of stress reactions. Neuroimage 119, 235–251.

67.

Nuño

, Gómez-Benito

, Carmona

, Pino

(2021) A systematic review of executive function and information processing speed in major depression disorder. Brain Sci 11, 147.

68.

Jungwirth

, Zehetmayer

, Hinterberger

, Kudrnovsky-Moser

, Weissgram

, Tragl

, Fischer

(2011) The influence of depression on processing speed and executive function in nondemented subjects aged 75. J Int Neuropsychol Soc 17, 822–831.

69.

Rutherford

, Choi

, Mass

, He

, O’Boyle

, Sneed

, Brown

, Brickman

, Wall

, Roose

(2021) Slowed processing speed disrupts patient expectancy in late life depression. Am J Geriatr Psychiatry 29, 619–630.

70.

Tops

, Boksem

(2011) A potential role of the inferior frontal gyrus and anterior insula in cognitive control, brain rhythms, and event-related potentials. Front Psychol 2, 330.

71.

Branger

, Arenaza-Urquijo

, Tomadesso

, Mézenge

, André

, de Flores

, Mutlu

, de La Sayette

, Eustache

, Chételat

, Rauchs

(2016) Relationships between sleep quality and brainvolume, metabolism, and amyloid deposition in late adulthood. Neurobiol Aging 41, 107–114.

72.

, McLeland

, Toedebusch

, Xiong

, Fagan

, Duntley

, Morris

, Holtzman

(2013) Sleep quality and preclinical Alzheimer disease. JAMA Neurol 70, 587–593.

73.

Ettore

, Bakardjian

, Solé

, Levy Nogueira

, Habert

, Gabelle

, Dubois

, Robert

, David

(2019) Relationships betweenobjectives sleep parameters and brain amyloid load in subjects atrisk for Alzheimer’s disease: The INSIGHT-preAD Study. Sleep 42, zsz137.

74.

Mander

, Marks

, Vogel

, Rao

, Lu

, Saletin

, Ancoli-Israel

, Jagust

, Walker

(2015) β-amyloid disrupts human NREM slow waves and related hippocampus-dependent memory consolidation. Nat Neurosci 18, 1051–1057.

75.

Sheffield

, Huang

, Rogers

, Blackford

, Heckers

, Woodward

(2021) Insula sub-regions across the psychosis spectrum: Morphology and clinical correlates. Transl Psychiatry 11, 346.