Insomnia might occur as result of increased cognitive and physiological arousal caused by acute or long acting stressors and associated cognitive rumination. This might lead to alterations in brain connectivity patterns as those captured by functional connectivity fMRI analysis, leading to potential insight about primary insomnia (PI) pathophysiology as well as the impact of long-term exposure to sleep deprivation. We investigated changes of voxel-wise connectivity patterns in a sample of 17 drug-naïve PI patients and 17 age-gender matched healthy controls, as well as the relationship between brain connectivity and age of onset, illness duration, and severity. Results showed a significant increase in resting-state functional connectivity of the bilateral visual cortex in PI patients, associated with decreased connectivity between the visual cortex and bilateral temporal pole. Regression with clinical scores originally unveiled a pattern of increased local connectivity as measured by intrinsic connectivity contrast (ICC), specifically resembling the default mode network (DMN). Additionally, age of onset was found to be correlated with the connectivity of supplementary motor area (SMA), and the strength of DMN←→SMA connectivity was significantly correlated with both age of onset (
Primary insomnia (PI) is a clinical condition characterized by troubles initiating or maintaining sleep, which is associated with daytime consequences and is not attributable to environmental circumstances or inadequate opportunity to sleep, as well as not to any other somatic or psychiatric cause [
Moreover, recent evidence has suggested that insomnia might have an impact on both night and day brain functioning, with changes in brain plasticity, assessed via transcranial magnetic stimulation (TMS), reported in patients with PI [
Therefore, we investigated differences in resting-state functional connectivity fMRI patterns in drug-naïve patients with PI compared to healthy controls, looking at correlations between insomnia-induced alterations and clinical variable, in particular age of onset and disease duration. Importantly, in order to avoid a priori selection of analysis ROIs/masks, we implemented a high-resolution FC analysis based on voxel-wise connectivity maps indexing both local and distributed functional connectivity patterns for each voxel in the brain. We hypothesized that PI patients will display altered connectivity in sensory systems and/or regions related to attention and memory processing. We also hypothesized that insomnia duration and age of onset might exert similar effects on brain functional connectivity patterns, with early onset possibly leading to a stronger disruption of physiological brain dynamics.
Seventeen drug-naïve insomnia patients and 17 age- and education-matched healthy controls participated in the study. Diagnosis was based on the ICSD-3 criteria for primary insomnia (PI). All the participants were right-handed (as measured using the Oldfield handedness scale), cognitively intact (Mini-Mental State Examination score > 24), and monolingual native speakers and underwent a general physical and neurological screening, as well as an assessment of their medical history. PI patients were diagnosed and enrolled at the Center for Sleep Medicine of the Le Scotte Hospital (Siena, Italy). Each patient completed self-report clinical scales assessing the severity of their sleep-related complaints (Pittsburgh Sleep Quality Index (PSQI), Insomnia Severity Scale (ISI)) and their mood status (Beck Depression Inventory (BDI)). Inclusion criteria for patients were as follows: (1) fist diagnosis of primary insomnia at our center; (2) 18 to 45 years old; (3) no evidence of other medical disorders, with particular reference to current or past neurological and psychiatric ones or other sleep disorders; (4) no history of assumption of drugs acting on the central nervous system; and (5) no previous treatment or diagnosis of primary insomnia. They were advised to drink no more than one cup of coffee (or two of tea) and to not assume any amount of alcohol or other type of drink with caffeine in the day of the radiological examination. Healthy controls showed a normal neurological exam, regular sleep-wake cycle and no sleep complaints. Exclusion criteria were as follows: (1) abnormalities in physical and neurological examination screening visit, (2) current or past substance abuse, (3) use of psychotropic medication within 3 months prior to inclusion, and (4) brain structural abnormalities at the magnetic resonance imaging (MRI) exam. All participants gave their written informed consent to the experimental procedure, which conformed to the Declaration of Helsinki. The study was approved by the local ethical committee.
Patients came to the Center for Sleep Medicine reporting sleep-related complaints involving “difficulty falling asleep or staying asleep, waking up early in the morning, and/or poor sleep quality with daytime consequences.” They were diagnosed for the first time by two neurologists (CDB and IS) licensed as Sleep Disorders Expert by the Italian Society for Sleep Medicine (Associazione Italiana Medicina del Sonno (AIMS);
The Pittsburgh Sleep Quality Index (PSQI) [
The Insomnia Severity Index (ISI) is a brief self-report measure assessing perception of the severity of sleep disturbance [
The Beck Depression Inventory (BDI) [
MRI data was acquired on a Philips Intera whole-body scanner. Resting-state fMRI data included 178 volumes with 33 axial slices covering the whole brain, acquired via a T2 BOLD-sensitive multislice echo planar imaging (EPI) sequence (TR/TE = 2.5 s/32 ms; field of view = 22 cm; image matrix = 64 × 64; voxel size = 3.44 × 3.44 × 3.8 mm3; flip angle = 75°). Structural imaging was performed using a whole brain T1-weighted Fast Field Echo 1 mm3 sequence (TR/TE = 30/4.6 ms, field of view = 250 mm, matrix 256 × 256, flip angle = 30°, slice number = 150, and scan time: 7 : 25 minutes). T2-weighted fluid-attenuated inverse recovery (FLAIR) images were also acquired to assess participants’ white matter integrity. Participants were provided with earplugs and were instructed to lay in the scanner with their eyes open, while fixating on a cross hair. They were asked to stay as still as possible. To monitor the patients’ state inside the scanner, the MRI technician monitored each patient via a camera placed inside the scanner for the entire fMRI acquisition. Particular care was taken to minimize head motion via vacuum cushions and custom-made padding.
fMRI data preprocessing and statistical analyses were carried out using SPM8 software (Statistical Parametric Mapping;
Individual connectivity maps were computed by means of the intrinsic connectivity contrast (ICC), a voxel-to-brain connectivity metric [
ICC is computed for each voxel in the brain, therefore producing a whole-brain map where the intensity of each voxel reflects the average
Resting state FC analysis was implemented using ad hoc scripts implemented in a Python and MATLAB computational environment, based on code and modules from the same software used for preprocessing of MRI data. Analysis was based on voxel-wise connectivity indexes using the intrinsic connectivity contrast (ICC) [
Voxel-wise connectivity maps were compared across PI patients and HC, using an analysis of covariance (ANCOVA) including age, gender, and BDI score as covariates. Results were considered significant at a threshold equal to
The same approach was used to derive patterns of disease-related modifications in patients’ connectivity profile. Analysis were run only in PI patients (
The selected PI patients (
Analysis of voxel-wise connectivity maps lead to significant differences in ICC patterns. Increase in connectivity of the bilateral occipital lobe was observed in PI patients with respect to controls (Figure
Voxel-wise FC changes. An increase in resting-state voxel-to-brain connectivity was observed in the occipital lobe of PI patients with respect to healthy controls. The seed-based connectivity profile of the significant occipital lobe cluster in (a) was compared across groups, highlighting an increase in local connectivity in PI patients and a decrease in connectivity in the bilateral temporal pole (b) (
Group differences in functional connectivity. Localization of the voxel-wise connectivity differences between PI patients and healthy controls are reported, with corresponding MNI coordinates. The results of seed-based analysis between the ICC cluster and the rest of the brain are also shown.
Procedure | Cluster MNI coordinates | Cluster localization | Cluster p-FDR | Increased/decreased connectivity | |||
---|---|---|---|---|---|---|---|
14 | −90 | 4 | 2328 | 973 voxels, primary visual cortex (left) | 0.00005 | ↑ | |
1076 voxels, primary visual cortex (right) | |||||||
341 voxels, lingual gyrus (right) | |||||||
132 voxels, lingual gyrus (left) | |||||||
38 | −66 | 12 | 2893 | 477 voxels, brain stem | 0.00003 | ↓ | |
153 voxels, frontal orbital cortex (right) | |||||||
139 voxels, hippocampus (right) | |||||||
132 voxels, temporal pole (right) | |||||||
111 voxels, lateral occipital cortex, inferior division (right) | |||||||
−42 | −48 | 0 | 1702 | 117 voxels, middle temporal gyrus, posterior division (left) | 0.0002 | ↓ | |
112 voxels, amygdala (left) | |||||||
92 voxels, hippocampus (left) | |||||||
80 voxels, parahippocampal gyrus, anterior division (left) | |||||||
51 voxels, middle temporal gyrus, temporooccipital part (left) | |||||||
0 | −86 | 10 | 590 | 156 voxels, intracalcarine cortex (left) | 0.008556 | ↑ | |
96 voxels, intracalcarine cortex (right) |
The regression model predicting age of onset highlighted a pattern of increased ICC in multiple clusters of voxels resembling the default mode network (DMN; Figures
Correlation with clinical scores. Results of the voxel-wise correlation between FC patterns and age of onset are shown in (a), highlighting a set of clusters closely resembling the default mode network (DMN), as confirmed by the seed-based connectivity profile of the same clusters computed on the entire study sample (b). Specifically, age of onset was positively correlated with the connectivity between the DMN cluster in (a) and the bilateral supplementary motor area (SMA) (c). By looking at resting-state activity in PI patients (c), SMA displays a negative correlation with the DMN in healthy controls (d), suggesting that chronic sleep deprivation might weaken such resting-state dynamic. Scatterplots (e) display individual FC strength between the DMN clusters in (a) and SMA in (c) as a function of age of onset and disease duration (
Correlation with age of onset. Results for both voxel-wise ICC and seed-based analysis are reported, with corresponding cluster size and coordinates in MNI space.
Procedure | Cluster MNI coordinates | Cluster localization | Cluster p-FDR | Increased/decreased connectivity | |||
---|---|---|---|---|---|---|---|
8 | 38 | 28 | 5609 | 759 voxels, paracingulate gyrus (right) | 0.00003 | ↑ | |
758 voxels, superior frontal gyrus (right) | |||||||
480 voxels, superior frontal gyrus (left) | |||||||
456 voxels, paracingulate gyrus (left) | |||||||
439 voxels, frontal pole (left) | |||||||
261 voxels, middle frontal gyrus (right) | |||||||
255 voxels, cingulate gyrus, anterior division | |||||||
243 voxels, frontal medial cortex | |||||||
125 voxels, middle frontal gyrus (left) | |||||||
8 | −56 | 38 | 5103 | 2996 voxels, precuneus cortex | 0.00004 | ↑ | |
1249 voxels, cingulate gyrus, posterior division | |||||||
122 voxels, cuneal cortex (right) | |||||||
107 voxels, intracalcarine cortex (right) | |||||||
44 | −68 | 26 | 1602 | 1220 voxels, lateral occipital cortex, superior division (right) | 0.000006 | ↑ | |
303 voxels, angular gyrus (right) | |||||||
−44 | −62 | 24 | 1264 | 1000 voxels, lateral occipital cortex, superior division (left) | 0.00007 | ↑ | |
228 voxels, angular gyrus (left) | |||||||
2 | 16 | 56 | 3671 | 1790 voxels, supplementary motor area (left) | 0.0002 | ↓ | |
1540 voxels, supplementary motor area (right) |
Data showed how chronic insomnia is able to induce changes in brain connectivity, with a specific impact on visual cortex resting-state activity. Moreover, individual differences in age of onset and insomnia duration were identified as a predictor of changes of connectivity patterns between the DMN and a core region of the motor system (SMA). Interestingly, age of onset displayed a significantly stronger correlation with fMRI alteration than disease duration, suggesting the importance of addressing insomnia-related effects on brain connectivity in younger adults to prevent long-lasting connectivity reshaping.
The most prominent difference in voxel-wise FC between PI patients and healthy controls was evident in the bilateral visual cortex. Interestingly, this finding has not been reported in any previous fMRI study on PI, whereas it fits with prior evidence of abnormal FC pattern within the occipital cortex in sleep-deprived healthy subjects [
In addition, seed-based analysis highlighted a reduction of connectivity between the occipital lobe and two clusters mapping on the bilateral temporal pole, in particular with the hippocampus. While modifications of temporal lobe activity have been reported in PI patients [
A significant correlation between individual connectivity patterns and both insomnia duration and age of onset was also found. We highlighted a very interesting correlation, yet preliminary and limited by sample size, between the age of onset and the strength of connectivity of regions highly resembling the DMN (i.e., medial prefrontal cortex, precuneus, and bilateral angular gyrus). Interestingly, using resting-state fMRI in healthy controls under controlled sleep deprivation, two studies have demonstrated an aberrant functional activity both within the DMN and between the DMN and its negatively correlated regions [
Age of onset-related remapping of brain functional architecture might be related to plasticity mechanisms, which seems to change across the lifespan [
If replicated in independent samples, the reversal of DMN-SMA dynamics highlighted in PI patients will suggest the need of early interventions aimed at counteracting such disruption of resting-state brain connectivity patterns, possibly using noninvasive brain stimulation (NIBS) [
The present findings suggest the importance of exploring the role of brain plasticity mechanism into compensating for early insomnia onset and prolonged exposure to sleep deprivation. Functional data also suggest a significant enhancement of resting-state activity in the visual cortex of PI patients, corroborating the hyperarousal theory of insomnia and possibly representing a target for therapeutic interventions.
All authors report no conflict of interest.