Hypersynchronized Magnetoencephalography Brain Networks in Patients with Mild Cognitive Impairment and Alzheimer's Disease in Down Syndrome

Abstract

Introduction:

The majority of individuals with Down syndrome (DS) show signs of Alzheimer's disease (AD) neuropathology in their fourth decade. However, there is a lack of specific markers for characterizing the disease stages while considering this population's differential features.

Methods:

Forty-one DS individuals participated in the study, and were classified into three groups according to their clinical status: Alzheimer's disease (AD-DS), mild cognitive impairment (MCI-DS), and controls (CN-DS). We performed an exhaustive neuropsychological evaluation and assessed brain functional connectivity (FC) from magnetoencephalographic recordings.

Results:

Compared with CN-DS, both MCI-DS and AD-DS showed a pattern of increased FC within the high alpha band. The neuropsychological assessment showed a generalized cognitive impairment, especially affecting mnestic functions, in MCI-DS and, more pronouncedly, in AD-DS.

Discussion:

These findings might help to characterize the AD-continuum in DS. In addition, they support the role of the excitatory/inhibitory imbalance as a key pathophysiological factor in AD.

Impact statement

The pattern of functional connectivity (FC) hypersynchronization found in this study resembles the largely reported Alzheimer's disease (AD) FC evolution pattern in population with typical development. This study supports the hypothesis of the excitatory/inhibitory imbalance as a key pathophysiological factor in AD, and its conclusions could help in the characterization and prediction of Down syndrome individuals with a greater likelihood of converting to dementia.

Introduction

The triplication of chromosome 21, causing Down's syndrome (DS), implies a genetic overexpression that gives rise to impaired neurogenesis, excessive astrocyte proliferation, and dendritic atrophy leading to impaired brain connectivity (Bartesaghi et al., 2011). This triplication prompts the overexpression of APP (amyloid precursor protein) and DYRKA1A (dual-specificity tyrosine-(Y)-phosphorylation regulated kinase 1A), both strongly related to Alzheimer's disease (AD) pathology (García-Cerro et al., 2017). Thus, it is not surprising that many people with DS develop AD as early as in their fourth decade of life (Cenini et al., 2012; Head et al., 2012 Sabbagh et al., 2011; Wilcock and Griffin, 2013). Importantly, recent findings show that neuropathological processes related to this disease could start even one decade before (Fortea et al., 2020).

In the sporadic variant of AD in individuals with typical development (TD), the dominant hypothesis assumes that either the soluble or aggregated form of the beta-amyloid (Aβ) protein leads to neuroinflammation, phosphorylation of tau protein, synaptic dysfunction, and, consequently, brain atrophy (Jack et al., 2010, 2013; Sperling et al., 2011; Vogel et al., 2020). Neurophysiological animal studies have shown that this proteinopathy causes an excitatory/inhibitory (E/I) imbalance, thereby affecting brain networks by causing hyperactivity during the first stages of the disease, hypoactivity in the middle stages, and the final collapse of the system (Busche and Konnerth, 2016; Busche et al., 2019).

These results have been partially validated in humans through the investigation of resting-state functional connectivity (FC) with electroencephalography or magnetoencephalography (MEG). TD individuals with mild cognitive impairment (MCI) show an increase in corticothalamic and anterior interhemispheric FC (Bajo et al., 2010; Cantero et al., 2009). In addition, resting-state evaluations revealed increased anteroposterior FC in TD individuals with MCI who converted to AD (López et al., 2014), in individuals who reported subjective cognitive decline (López-Sanz et al., 2017), and in first-degree relatives of AD patients (Ramírez-Toraño et al., 2021). In contrast, TD AD patients showed reduced FC during resting state, mainly in the alpha and beta bands (Pusil et al., 2019), with the magnitude of this reduction being related to both the severity of the disease (Stam et al., 2006) and a progressive loss of brain functional hubs (Yu et al., 2017).

Studies of FC in DS, however, have focused on subjects without MCI or AD. Compared with TD participants (average age: 46.95 years), Vega and colleagues (2015) found that adults with DS (average age 39 years) presented higher levels of between-network FC, whereas the within-network FC seemed preserved, a pattern that could be consistent with AD-related pathology. Wilson and colleagues (2019) found alterations in the default mode network (DMN) in the DS population compared with matched TD controls, even in the absence of Aβ pathology. Importantly, FC anomalies were even more extensive in DS participants showing Aβ pathology, suggesting a relationship between amyloid deposition and damaged DMN connectivity. This alteration in the DMN could be, according to the authors, a marker of the pre-clinical neurodegeneration (Wilson et al., 2019). Indeed, studies of amyloid accumulation in TD participants show that the deposition pattern highly matches the areas involved in the DMN (Buckner et al., 2005, 2009), suggesting a tight relationship between Aβ—and, thus, AD—pathology, and FC alterations.

Notably, there are no FC studies performed in DS population with MCI or AD. Considering this framework, this study aimed to investigate FC in a DS cohort at three stages of the AD continuum: participants without any clinical symptoms, subjects in the predementia stage, and participants in the dementia stage. To the best of our knowledge, this is the first MEG study measuring FC along the AD continuum in the DS population. The already described pieces of evidence point to a similar neuropathological background in TD and DS populations. Therefore, we hypothesize that FC patterns in MCI-DS and AD-DS cases will mirror typical neurophysiological findings already described for TD individuals with AD.

Materials and Methods

Participants

Forty-one individuals with DS participated in the study. They all had genetically confirmed karyotype; cases with mosaicism and translocations were excluded. All subjects presented a mild or moderate level of intellectual disability according to Diagnostic and Statistical Manual of Mental Disorders-5 criteria (American Psychiatric Association, 2013).

The sample was divided into three groups: (1) DS individuals who did not match the criteria for MCI or AD diagnosis (CN-DS group), (2) DS subjects who fit the criteria for MCI (MCI-DS group), and (3) DS individuals who matched the criteria for AD (AD-DS group). Participants were not receiving any drug treatment that could interfere with MEG or neuropsychological assessment. Cases with clinical hypo/hyperthyroidism, uncontrolled B₉/B₁₂ vitamin deficiency, delirium, disorders that could show similar symptoms than cognitive impairment (e.g., depression), and severe uncorrected sensory impairment (auditory or visual) were excluded.

Candidates were recruited from the Unit for Attention of Adults with Down Syndrome (La Princesa University Hospital, Madrid, Spain). The study was conducted following The Code of Ethics of the World Medical Association (i.e., Declaration of Helsinki), and the Clinical Research Ethical Committee of La Princesa University Hospital approved the protocol. In addition to receiving verbal or written permission from DS cases, we obtained written informed consent from all our participants' parents or legal guardians. The demographic characteristics of the sample are displayed in Table 1.

Table 1.

Demographic Characteristics and Neuropsychological Performance

	CN-DS (n = 14)	MCI-DS (n = 14)	AD-DS (n = 13)
Age^*	44.64 ± 3.30	51.64 ± 3.95	53.54 ± 6.58
Gender	4 M—10 F	8 M—6 F	4 M—9 F
Abstract thinking (CAMCOG-DS^*)	2.57 ± 2.59/0	0.54 ± 1.66/1	1.00 ± 1.95/1
Total score (CAMCOG-DS^*)	76.64 ± 17.86/0	59.08 ± 15.37/1	56.92 ± 15.91/1
Verbal learning (BT-ID^*)	23.29 ± 10.56/0	18.31 ± 10.19/1	10.00 ± 6.45/0
Working memory (BRIEF^*)	15.58 ± 3.09/2	18.33 ± 4.31/2	21.09 ± 5.07/2
Temporal orientation (BT-ID^*)	47.31 ± 28.35/1	15.54 ± 20.10/1	6.85 ± 8.44/0
Working memory I (BT-ID^*)	4.43 ± 3.88/0	0.69 ± 1.18/1	1.46 ± 3.57/0
Delayed visual memory (BT-ID^*)	0.86 ± 1.10/0	4.36 ± 4.20/3	7.58 ± 4.32/1

Mean ± SD/number of participants not completing the protocol.

AD-DS, Alzheimer's disease-Down syndrome group; BRIEF, Behavior Rating Inventory of Executive Function; BT-ID, Barcelona Test-Intellectual Disability; CAMCOG-DS, Cambridge Cognitive Examination for Older Adults with Down's Syndrome; CN-DS, Control-Down's syndrome group; F, female; M, male; MCI-DS, mild cognitive impairment-Down's syndrome group; SD, standard deviation.

Clinical assessment

Assessment for all participants was performed using a comprehensive clinical and neuropsychological protocol. First, the diagnosis of MCI and AD was based on expert clinical judgment, as it is the standard recommendation for DS (Fenoll et al., 2017; Krinsky-McHale and Silverman, 2013; Pujol et al., 2018; Sheehan et al., 2015). In detail, a diagnosis of MCI was established based on (1) a report of cognitive impairment either by a reliable informant or by the patient (confirmed by a reliable informant) implying a change from previous capacities and (2) a proof of the absence of clinically relevant deterioration in adaptive skills (García-Alba et al., 2019). The diagnosis of AD was established when the patient previously met MCI criteria and showed a noticeable decline in adaptive skills associated with memory impairment, besides, at least one of the following disorders had to be present: aphasia, apraxia, agnosia, or dysexecutive syndrome (Fenoll et al., 2017; García-Alba et al., 2019; Pujol et al., 2018).

The neuropsychological tests used were especially sensitive to mental deterioration in DS and adapted to the Spanish-speaking population (Esteba-Castillo et al., 2013, 2017). Specifically, the protocol was composed of(1) the Cambridge Cognitive Examination for Older Adults with Down's Syndrome (CAMCOG-DS) Spanish version (Esteba-Castillo et al., 2013; original version: Ball et al., 2006), whose abstract thinking (AT) and total score (TS) subtests were used; (2) temporal orientation (TO), delayed visual memory (DVM), verbal learning-verbal short-term memory (VL_VSTM), and working memory I (WM_I) subtests of the Barcelona Test-Intellectual Disability (original version: Peña-Casanova, 2005); (3) the working memory (WM) subtest of the Behavior Rating Inventory of Executive Function-Parents (Gioia et al., 2010); and (4) the Cambridge Examination for Mental Disorders of Older People with Down's Syndrome and Others with Intellectual Disabilities (CAMDEX-DS-Spanish version: Esteba-Castillo et al., 2013; original version: Ball et al., 2006), which was only used as a diagnostic tool.

The average results per group in the tests are presented in Table 1. For the sake of clarity, only the relevant results are displayed here. The complete information on neuropsychological performance is presented in Supplementary Table S1.

Magnetoencephalography

Data acquisition

MEG data were acquired using an Elekta Vectorview (MEGIN OY, Helsinki, FI) magnetometer with 306 sensors (102 magnetometers and 204 planar gradiometers) covering the whole head. The scanning array was located inside a magnetically shielded room (Vacuumschmelze GmbH, Hanau, DE) in the Center for Biomedical Technology (Madrid, Spain). Before the acquisition, participants' setup was performed by placing four head position indicator (HPI) coils on the head, as well as two bipolar electrodes above and below the left eye aimed to capture blinks and eye movements. Lastly, we acquired a set of head shape points, plus the location of the four HPI coils, using a FASTRAK digitizer (Polhemus, Colchester, VT).

Four minutes of eye-closed resting-state MEG data were acquired. During the acquisition, the participant was instructed to be still and relaxed while staying awake. Besides, the HPI coils allow us to track the head movements. Data were band pass filtered online in the 0.1 to 330 Hz band, and then digitized using a sample rate of 1000 Hz. Before preprocessing, we applied a spatial filter to the MEG signal using the temporal extension of the signal space separation (tSSS) method (Taulu and Simola, 2006) as implemented by MaxFilter software (version 2.2; Elekta Oy, Helsinki, FI), using a correlation threshold of 0.90 and a time window of 10 sec. We also used tSSS to compensate for the head movements using its MaxMove extension.

Data analysis

MEG data were preprocessed using Matlab (The Mathworks, Natick, MA) and FieldTrip toolbox (Oostenveld et al., 2011). First, an MEG expert marked the artifacts; then, we segmented the remaining artifact-free data into nonoverlapping epochs of 4 sec. We only used data from participants with at least 20 clean segments to obtain meaningful results. Owing to the high redundancy of the data after temporal signal space separation (Garcés et al., 2017), we used only the magnetometers for the subsequent analysis.

Owing to the inherent issues encountered while working with this population in noisy environments, performing the magnetic resonance imaging scan was not possible for most participants. Thus, we decided to use standard anatomy based on the Montreal Neurological Institute (MNI) template (Douw et al., 2017) for MEG source reconstruction. As a source model, we defined a homogeneous grid in MNI space, with an even separation of 10 mm between source positions, labeling each source according to a reduced version of the Harvard–Oxford probabilistic atlas (Desikan et al., 2006), resulting in a total of 1485 source positions in 64 cortical areas. The MNI template was segmented into its different tissues using the unified segmentation algorithm (Ashburner and Friston, 2005). We generated a single shell head model defined by the inner skull boundary (the union of gray matter, white matter, and cerebrospinal fluid). We also produced a surface defining the scalp, and then linearly transformed it to match the individual head shape points. This linear transformation was applied to the source model and head model, obtaining a participant-specific forward model.

As an inverse model, we used a linearly constrained minimum variance beamformer (Van Veen et al., 1997). The beamformer spatial filter was calculated separately for low alpha (8–10 Hz) and high alpha (10–12 Hz) bands. We used a 1800th order finite impulse response filter designed with a Hamming window, a two-passed procedure, and 2 sec of real data at each side as padding. Together with the band-specific spatial filter, these band pass filtered data were used to reconstruct the source space time series for each source position.

We estimated the FC between every two sources using the amplitude envelope correlation (AEC; Hipp et al., 2012). To avoid ghost synchronizations due to leakage, we first orthogonalized the instantaneous time series using a regression-based approach (Brookes et al., 2011), and extracted the instantaneous amplitude using the Hilbert analytical signal. To avoid edge artifacts, we padded the segments using 2 sec of real data before applying the Hilbert filter. We then calculated the AEC for each epoch using Pearson's correlation coefficient and averaged the result across epochs. Finally, the FC between each pair of areas in the Harvard–Oxford atlas was estimated as the average of all the AEC values between one source position in the first area and one source position in the second area. The final result was a set of two 64 by 64 FC matrices, one for each frequency band.

Statistical analysis

Distribution of genders across groups was compared using a chi-squared test. The age distribution was analyzed through an analysis of variance (ANOVA) test, taking the significant results to be further studied by applying Fisher's least difference test. For the cognitive tests, we used a nonparametric Kruskal–Wallis statistic, correcting for multiple comparisons with a false discovery rate (FDR; Benjamini and Hochberg, 1995) of 10%, and further studying the significant results by using Fisher's least difference test.

The differences in FC were studied using an analysis of covariance (ANCOVA) test, including the effect of age as a covariate, as significant differences across groups were found concerning this variable. As the FC values did not show a normal distribution, we used a nonparametric approach based on 50,000 random permutations of the original data. The reported p-value is based on the resulting null distribution. We then corrected the results for multiple comparisons using an FDR of 10%. The significant results were further analyzed using a nonparametric post hoc test based on the same set of permutations. All statistical analyses were performed using in-house MATLAB scripts.

Results

Sample characteristics

The ANOVA test indicated that the effect of age was significant (p < 0.0001). Post hoc analyses showed that the CN-DS participants were significantly younger than the MCI-DS (p = 0.0011) and the AD-DS (p < 0.0001) participants. The complete p values and size effects are given in Supplementary Table S2.

Neuropsychological performance

Seven neuropsychological tests showed significant between-group differences. For the sake of clarity, we only present the significant p values and effect sizes in Table 2, whereas the complete set of results are given in Supplementary Table S3.

Table 2.

Statistics, p Values, and Size Effects Associated with Neuropsychological Performance

	General	CN-DS vs. MCI-DS	CN-DS vs. AD-DS	MCI-DS vs. AD-DS
Abstract thinking (CAMCOG-DS)
K	75.690
p value	0.0227	0.0094	0.0429	NS
Size effects	0.1466
Total score (CAMCOG-DS)
K	92.439
p value	0.0098	0.0122	0.0070	NS
Size effects	0.1906
Verbal learning (BT-ID)
K	113.784
p value	0.0034	NS	0.0008	0.0358
Size effects	0.2468
Working memory I (BT-ID)
K	73.682
p value	0.0251	0.0249	0.0155	NS
Size effects	0.1413
Delayed visual memory (BT-ID)
K	161.946
p value	0.0003	0.0480	0.0001	NS
Size effects	0.3735
Working memory (BRIEF)
K	71.763
p value	0.0276	NS	0.0075	NS
Size effects	0.1362
Temporal orientation (BT-ID)
K	128.644
p value	0.0016	0.0045	0.0009	NS
Size effects	0.2859

In the post hoc comparison, only p values <0.05 are presented; otherwise, the result is marked as NS.

K, Kruskal–Wallis statistic; NS, not significant.

In detail, the MCI-DS group exhibited significantly weaker performance than the CN-DS group in AT, TS, WM, DVM, and TO. As expected, the AD-DS group showed a significantly weaker execution in all tests than the CN-DS group. Notably, VL_VSTM was the only variable wherein significant differences between the MCI-DS and AD-DS groups were found, so that AD-DS performed worse than MCI-DS. Of note is the high standard deviations found in the majority of the tests, due to the reduced number of participants, which could be influenced by the loss of some subjects, but also by a phenomenon already well known in DS, high cognitive variability (García-Alba et al., 2017; Wilson et al., 2019; Table 1).

Functional connectivity

The ANCOVA (taking into account participants' age) showed a set of seven anteroposterior links wherein the FC values differed significantly across groups. The most relevant differences emerged in the links between the left superior frontal gyrus and several posterior regions such as the left precuneus, the right cuneus, and both left and right occipital poles. Post hoc tests revealed that both the AD-DS and the MCI-DS groups presented higher synchronization values in all these links compared with the CN-DS group. Table 3 gives the average connectivity values within the significant links across groups and the significance of the ANCOVA test for all of them.

Table 3.

Analysis of Covariance Test Results

Links	Functional connectivity (mean ± SD)			p
Links	CN-DS	MCI-DS	AD-DS	p
Average	0.0029 ± 0.0962	0.1072 ± 0.0583	0.0731 ± 0.1270
lSFG-rSTG-p	−0.0013 ± 0.1133	0.1379 ± 0.0918	0.1003 ± 0.1380	0.0002
lSFG-rCU	−0.0012 ± 0.1356	0.1144 ± 0.0796	0.0873 ± 0.1325	0.0001
lSFG-rOP	0.0060 ± 0.1002	0.0938 ± 0.0568	0.0579 ± 0.1344	0.0001
lSFG-lOP	0.0070 ± 0.1065	0.0923 ± 0.0691	0.0714 ± 0.1281	0.0001
lSFG-lPc	−0.0108 ± 0.1260	0.1016 ± 0.0709	0.0749 ± 0.1412	<0.0001
lSFG-lLin	0.0215 ± 0.1033	0.1058 ± 0.0623	0.0633 ± 0.1308	0.0001
lLin-lIFG	−0.0006 ± 0.1034	0.1043 ± 0.0862	0.0567 ± 0.1258	<0.0001

lIFG, left inferior frontal gyrus; lLin, left lingual gyrus; lPc, left precuneus; lOP, left occipital pole; lSFG, left superior frontal gyrus; rCU, right cuneus; rOP, right occipital pole; rSTG-p, right posterior superior temporal gyrus.

We depict the FC results in Figure 1, where the top panel presents the brain areas that showed significant differences in FC, and the bottom panel displays the average FC of each subject across the noteworthy links and the average FC of each group. Similar graphs, but segregated by association, are presented in Supplementary Figure S1. In addition, the effect of the significance threshold is presented in Supplementary Figure S2.

FIG. 1.

Significant differences in functional connectivity. The top panel shows the areas with substantial differences in the ANCOVA test. Red indicates that the FC in MCI-DS and AD-DS groups is significantly higher than in the CN-DS group. The bottom panel presents the average FC values of each subject across the significant links. The big dot represents the mean of the group. AD-DS, Alzheimer's disease-Down syndrome group; ANCOVA, analysis of covariance; CN-DS, Control-Down's syndrome group; FC, functional connectivity; MCI-DS, mild cognitive impairment-Down's syndrome group.

Discussion

In this study, we evaluated FC in adults with DS at different stages along the AD continuum. Our findings show increased FC in the high alpha band in both MCI-DS and AD-DS compared with CN-DS. In addition, the neuropsychological evaluation showed a generalized cognitive decline in the MCI-DS group, especially in mnesic functions, a pattern that was strengthened in the AD-DS group. Our results support the exploration of network disruption to track the stages of the AD continuum.

Cognitive processes are supported in the brain by large scale connections between distinct regions (Andrews-Hanna et al., 2010; Sporns, 2011; Thomas Yeo et al., 2011) that dynamically engage in dealing with internal and external information (Fox et al., 2005; Greicius et al., 2003; Singh and Fawcett, 2008). Therefore, an aberrant connectivity pattern, either by excess or deficiency of activation, may indicate network malfunction (Schnitzler and Gross, 2005; Uhlhaas and Singer, 2006). Previous research has consistently found an increase in alpha band synchronization in healthy older adults positive for amyloid-positron emission tomography (PET) (Nakamura et al., 2017), MCI patients who later develop AD dementia (López et al., 2014), individuals reporting subjective cognitive decline (López-Sanz et al., 2017), and first-degree relatives of AD patients (Ramírez-Toraño et al., 2021). In addition, some of these reports showed that higher synchronization was related to poorer neuropsychological performance involving WM, executive functioning, and language (López et al., 2014; López-Sanz et al., 2017). The pattern of brain alterations found in this study goes in line with these previous findings, concluding that the AD continuum presents similar courses and FC manifestations in both DS and TD individuals.

Thus, as an essential inference of the current investigation, we suggest that the level of participants' hypersynchrony could be used to track their stage across the AD continuum. Pusil and colleagues (2019) followed for 24 months a cohort of MCI patients, some of whom converted to AD dementia during the study. Participants' baseline MEG data were acquired when all participants were diagnosed as MCI; MEG acquisition was repeated in a follow-up session 2 years after the baseline time point. At baseline, MCI patients who later converted to AD showed increased FC in the beta band in comparison with nonconverters. However, at the time of the dementia diagnosis, these patients showed reduced synchrony, indicating a communication breakdown. In contrast, the stable/nonconverter MCI patients showed increased FC at follow-up compared with baseline, suggesting a worsening of their pathological status despite not qualifying for dementia diagnosis. This pattern of brain network behavior across time was named the “X” model, with each arm of the “X” reflecting the trajectory of FC in either stable (the arm going upward) or progressive (the arm going downward, leading to AD conversion) MCI status. Since the same profile seems to appear in DS, MCI-DS individuals showing higher FC would be expected to progress and qualify to be considered as AD-DS in the following years. However, this hypothesis should be validated with a longitudinal follow-up of our sample.

According to the E/I imbalance phenomenon (Busche and Konnerth, 2016), the synchronization anomalies already mentioned could be explained as caused by the presence of different forms of amyloid. Studies on the microarchitecture of the cerebral cortex in humans with AD have shown a loss of GABAergic terminals in the vicinity of amyloid plaques (Garcia-Marin et al., 2009). Within this framework, the presence of amyloid aggregates in the brain would increase the cortical excitability, which, in turn, would increase the likelihood of observing altered synchronization between neuronal populations during MEG recordings (Canuet et al., 2015; Nakamura et al., 2017). Interestingly, when DS participants were divided into amyloid-PET positive and negative, the former showed more aberrant dysfunction on the posterior DMN nodes. This difference among DS individuals could be due to a more advanced neuropathology status in the amyloid-PET positive group, which would mirror the posterior hubs' loss in demented patients (Yu et al., 2017).

All these pieces of evidence seem to indicate that people with DS are more prone to AD pathology than TD populations, or that the neuropathology might be present even in the absence of AD symptomatology. However, not all individuals with AD pathology will develop the disease (Mullins et al., 2013), and the understanding of this lack of clinical manifestations in the presence of neuropathology is crucial. From birth, people with DS present varying degrees of intellectual disability. This cognitive deficit can “mask” the diagnosis of cognitive impairment or dementia. The neuropsychological tests applied in this study exhibit high sensitivity to the neuropsychological changes made in DS at different AD continuum stages. Our results show a generalized cognitive impairment in the MCI-DS group measured with the CAMCOG-DS TS, and this impairment is accentuated in the AD-DS group. Also, MCI-DS individuals present an amnesic profile affecting DVM and verbal working memory, specifically, whereas short-term verbal learning ability (VL_VSTM) is preserved in this group but lost in the AD-DS group. A temporary disorientation pattern was observed both in the MCI-DS and the AD-DS groups, as expected. The AD-DS group showed, as well, a loss of ability to consolidate verbal and visual information, with difficulties not only in coding but also in retrieving the given material. Besides, these patients presented deterioration in memory processes involved in executive functions. These findings are in line with previous reports that defined AD's clinical onset in DS by using its cognitive semiology (Firth et al., 2018). This previous research suggested that changes in memory and orientation might be crucial in detecting progression to AD. Still, these behavioral markers may be masked by the cognitive profile inherent to DS phenotype and, thus, changes could only be detected by highly sensitive neuropsychological tests (Dekker et al., 2018; Esteba-Castillo et al., 2013, 2017; García-Alba et al., 2017). Hence, our combined neurophysiological and neuropsychological findings constitute a clearer definition of the profile of subjects with DS along the AD continuum, which is of utmost clinical relevance. Specifically, the characterization of MCI in DS can help to complement and improve clinical diagnoses, differentiating, as well, the early stages of the disease from the more advanced stages. Furthermore, our results suggest that FC can be an ideal diagnostic tool, supporting previous descriptions of increased FC being associated with lower communications skills (Pujol et al., 2015) and significantly reduced memory performance (Alemany González, 2019) in people with DS.

Nevertheless, some limitations should be acknowledged for the current investigation. First, the initial sample of 41 individuals resulted relatively small after the three-group division, which may have led to a relative loss in statistical power and, thus, to miss some interesting effects. Second, we were forced to use a T1-MRI template instead of the individual structural data, what could diminish the spatial resolution in the source reconstruction process. However, this procedure has already been used by our team in previous studies (García-Alba et al., 2019). Also, similar problems were found in earlier studies examining the relationship between AD and DS using functional magnetic resonance imaging, where the participants' excessive movement led to many experimental difficulties (Pujol et al., 2014). The eventual loss of spatial resolution brought by using a template head model was overcome using an analysis based on parcellation, as the results show high spatial overlap with previous literature. Furthermore, our study's primary intention was to find an MEG-FC profile that could track the group differences across the AD stages and not to fully characterize the AD network traits in DS.

Conclusion

This study provides new and fundamental findings that could help better understand the process of dementia in DS subjects in two main ways. First, we suggest a new framework for the characterization and prediction of DS individuals having a greater likelihood of converting to dementia by using the “X” model (Pusil et al., 2019). Second, the pattern of functional hypersynchronization found in this study supports the role of the E/I imbalance as a key pathophysiological factor in AD, which could inspire new pharmacological strategies of intervention, such as the use of antiepileptic drugs or GABAergic modulators. Future studies should assess the DS population applying a longitudinal design, to further evaluate the predictions observed here, as well as to compare the predictive ability of FC with that of other current biomarkers.

Footnotes

Acknowledgments

The authors thank the participation of patients and their families in the effort to attend all medical appointments necessary for the study.

Authors' Contributions

F.M. and A.F. conceptualized the study and obtained the funding for it. J.G.A. and S.E.C. recruited the sample and collected the data. F.R.T. and R.B. performed the analysis. F.R.T., J.G.A., R.B., S.E.C., and L.V. worked in the writing of the article and the interpretation of the findings. E.P., F.M., and A.F. reviewed the interpretation of the findings. L.V. reviewed the writing style and helped communicate the findings.

Author Disclosure Statement

The authors report no competing interests. The funding sources had no role in the study design, data collection, data analyses, or data interpretation.

Funding Information

F.R.T. was supported by a predoctoral grant by the Spanish Ministry of Economy and Competitiveness (BES-2016-076869). E.P. received financial support of the same ministry through the grants TEC2016-80063-C3-2-R and PID2019-111537GB-C22. L.V. was supported by a postdoctoral fellowship within the “Program to attract talented researchers to include them in research groups in Madrid,” Type 2 (2018-T2/BMD-10991). This study was funded by the Ministry of Economy and Competitiveness (Institute of Health Carlos III; PI12/02019; Spain) and the Jérôme Lejeune Foundation (France).

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

References

Alemany González

. 2019. Neural correlates of memory impairment and rescue in a mouse model of down syndrome [Thesis in Spanish]. Barcelona, Spain: Universitat Pompeu. Fabra.

American Psychiatric Association. 2013. Diagnostic and Statistical Manual of Mental Disorders. Washington, D.C.: American Psychiatric. Association.

Andrews-Hanna

, Reidler

, Sepulcre

, et al. 2010. Functional-anatomic fractionation of the brain's default network. Neuron, 65:550–562.

Ashburner

, Friston

. 2005. Unified segmentation. NeuroImage, 26:839–851.

Bajo

, Maestú

, Nevado

, et al. 2010. Functional connectivity in mild cognitive impairment during a memory task: implications for the disconnection hypothesis. J Alzheimers Dis, 22:183–193.

Ball

, Holland

, Huppert

, et al. 2006. CAMDEX-DS; The Cambridge Examination for Mental Disorders of Older People with Down's Syndrome and Others with Intellectual Disabilities. Cambridge, UK: Cambridge University. Press.

Bartesaghi

, Guide

, Ciani

. 2011. Is it possible to improve neurodevelopmental abnormalities in Down syndrome?. Rev Neurosci, 22:419–455.

Benjamini

, Hochberg

. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B (Methodological), 57:289–300.

Brookes

, Hale

, Zumer

, et al. 2011. Measuring functional connectivity using MEG: methodology and comparison with fcMRI. NeuroImage, 56:1082–1104.

10.

Buckner

, Sepulcre

, Talukdar

, et al. 2009. Cortical hubs revealed by intrinsic functional connectivity: mapping, assessment of stability, and relation to Alzheimer's disease. J Neurosci, 29:1860–1873.

11.

Buckner

, Snyder

, Shannon

, et al. 2005. Molecular, structural, and functional characterization of Alzheimer's disease: evidence for a relationship between default activity, amyloid, and memory. J Neurosci, 25:7709–7717.

12.

Busche

, Konnerth

. 2016. Impairments of neural circuit function in Alzheimer's disease. Philos Trans R Soc B Biol Sci, 371:20150429.

13.

Busche

, Wegmann

, Dujardin

, et al. 2019. Tau impairs neural circuits, dominating amyloid-β effects, in Alzheimer models in vivo. Nat Neurosci, 22:57–64.

14.

Cantero

, Atienza

, Gomez-Herrero

, et al. 2009. Functional integrity of thalamocortical circuits differentiates normal aging from mild cognitive impairment. Hum Brain Mapp, 30:3944–3957.

15.

Canuet

, Pusil

, López

, et al. 2015. Network disruption and cerebrospinal fluid amyloid-beta and phospho-tau levels in mild cognitive impairment. J Neurosci, 35:10325–10330.

16.

Cenini

, Dowling

ALS

, Beckett

, et al. 2012. Association between frontal cortex oxidative damage and beta-amyloid as a function of age in Down syndrome. Biochim Biophys Mol Basis Dis, 1822:130–138.

17.

Dekker

, Sacco

, Carfi

, et al. 2018. The Behavioral and Psychological Symptoms of Dementia in Down Syndrome (BPSD-DS) Scale: comprehensive assessment of psychopathology in down syndrome. J Alzheimers Dis, 63:797–819.

18.

Desikan

, Ségonne

, Fischl

, et al. 2006. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage, 31:968–980.

19.

Douw

, Nieboer

, Stam

, et al. 2017. Consistency of magnetoencephalographic functional connectivity and network reconstruction using a template versus native MRI for co-registration. Hum Brain Mapp, 119:104–119.

20.

Esteba-Castillo

, Dalmau-Bueno

, Ribas-Vidal

, et al. 2013. Adaptación y validación del Cambridge Examination for Mental Disorders of Older People with Down's Syndrome and Others with Intellectual Disabilities (CAMDEX-DS) en población española con discapacidad intelectual. Rev. Neurol.

21.

Esteba-Castillo

, Peña-Casanova

, García-Alba

, et al. 2017. Barcelona test for intellectual disability: a new instrument for the neuropsychological assessment of adults with intellectual disability. Rev Neurol, 64:433–444.

22.

Fenoll

, Pujol

, Esteba-Castillo

, et al. 2017. Anomalous white matter structure and the effect of age in down syndrome patients. J Alzheimers Dis, 57:61–70.

23.

Firth

, Startin

, Hithersay

, et al. 2018. Aging related cognitive changes associated with Alzheimer's disease in Down syndrome. Ann Clin Transl Neurol, 5:741–751.

24.

Fortea

, Vilaplana

, Carmona-Iragui

, et al. 2020. Clinical and biomarker changes of Alzheimer's disease in adults with Down syndrome: a cross-sectional study. Lancet, 395:1988–1997.

25.

Fox

, Snyder

, Vincent

, et al. 2005. The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proc Natl Acad Sci USA, 102:9673–9678.

26.

Garcés

, López-Sanz

, Maestú

, et al. 2017. Choice of magnetometers and gradiometers after signal space separation. Sensors, 17:2926.

27.

García-Alba

, Esteba-Castillo

, Castellanos López

MÁ

, et al. 2017. Validation and normalization of the tower of London-Drexel University Test 2 nd edition in an adult population with intellectual disability. Spanish J Psychol, 20:E32.

28.

García-Alba

, Ramírez-Toraño

, Esteba-Castillo

, et al. 2019. Neuropsychological and neurophysiological characterization of mild cognitive impairment and Alzheimer's disease in Down syndrome. Neurobiol Aging, 84:70–79.

29.

García-Cerro

, Rueda

, Vidal

, et al. 2017. Normalizing the gene dosage of Dyrk1A in a mouse model of Down syndrome rescues several Alzheimer's disease phenotypes. Neurobiol Dis, 106:76–88.

30.

Garcia-Marin

, Blazquez-Llorca

, Rodriguez

, et al. 2009. Diminished perisomatic GABAergic terminals on cortical neurons adjacent to amyloid plaques. Front Neuroanat, 3:28.

31.

Gioia

, Isquith

, Guy

, et al. 2010. Child neuropsychology TEST REVIEW behavior rating inventory of executive function. Child Neuropsychol, 6:235–238.

32.

Greicius

, Krasnow

, Reiss

, et al. 2003. Functional connectivity in the resting brain: a network analysis of the default mode hypothesis. Proc Natl Acad Sci USA, 100:253–258.

33.

Head

, Silverman

, Patterson

, et al. 2012. Editorial aging and down syndrome. Curr Gerontol Geriatr Res, 2012:412536.

34.

Hipp

, Hawellek

, Corbetta

, et al. 2012. Large-scale cortical correlation structure of spontaneous oscillatory activity. Nat Neurosci, 15:884–890.

35.

Jack

, Knopman

, Jagust

, et al. 2010. Hypothetical model of dynamic biomarkers of the Alzheimer's pathological cascade. Lancet Neurol, 9:119–128.

36.

Jack

, Knopman

, Jagust

, et al. 2013. Tracking pathophysiological processes in Alzheimer's disease: an updated hypothetical model of dynamic biomarkers. Lancet Neurol, 12:207–216.

37.

Krinsky-McHale

, Silverman

. 2013. Dementia and mild cognitive impairment in adults with intellectual disability: issues of diagnosis. Dev Disabil Res Rev, 18:31–42.

38.

López

, Bruña

, Aurtenetxe

, et al. 2014. Alpha-band hypersynchronization in progressive mild cognitive impairment: a magnetoencephalography study. J Neurosci, 34:14551–14559.

39.

López-Sanz

, Bruña

, Garcés

, et al. 2017. Functional connectivity disruption in subjective cognitive decline and mild cognitive impairment: a common pattern of alterations. Front Aging Neurosci, 9:109.

40.

Mullins

, Daly

, Simmons

, et al. 2013. Dementia in Down's syndrome: an MRI comparison with Alzheimer's disease in the general population. J Neurodev Dis, 5:19.

41.

Nakamura

, Cuesta

, Kato

, et al. 2017. Early functional network alterations in asymptomatic elders at risk for Alzheimer's disease. Sci Rep, 7:6517.

42.

Oostenveld

, Fries

, Maris

, et al. 2011. FieldTrip: open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data. Comput Intell Neurosci, 2011:156869.

43.

Peña-Casanova

. 2005. Programa integrado de exploración neuropsicológica: TRB: Test Barcelona Revisado, 11:66–79.

44.

Pujol

, del Hoyo

, Blanco-Hinojo

, et al. 2015. Anomalous brain functional connectivity contributing to poor adaptive behavior in Down syndrome. Cortex, 64:148–156.

45.

Pujol

, Fenoll

, Ribas-Vidal

, et al. 2018. A longitudinal study of brain anatomy changes preceding dementia in Down syndrome. NeuroImage, 18:160–166.

46.

Pujol

, Macià

, Blanco-Hinojo

, et al. 2014. Does motion-related brain functional connectivity reflect both artifacts and genuine neural activity?. NeuroImage, 101:87–95.

47.

Pusil

, López

, Cuesta

, et al. 2019. Hypersynchronization in mild cognitive impairment: the ‘X’ model. Brain, 142:3936–3950.

48.

Ramírez-Toraño

, Bruña

, De Frutos-Lucas

, et al. 2021. Functional connectivity hypersynchronization in relatives of Alzheimer's disease patients: an early E/I balance dysfunction?. Cereb Cortex, 31:1201–1210.

49.

Sabbagh

, Fleisher

, Chen

, et al. 2011. Positron emission tomography and neuropathologic estimates of fibrillar amyloid-β in a patient with down syndrome and Alzheimer disease. Arch Neurol, 68:1461–1466.

50.

Schnitzler

, Gross

. 2005. Normal and pathological oscillatory communication in the brain. Nat Rev Neurosci, 6:285–296.

51.

Sheehan

, Sinai

, Bass

, et al. 2015. Dementia diagnostic criteria in Down syndrome. Int J Geriatr Psychiatry, 30:857–863.

52.

Singh

, Fawcett

. 2008. Transient and linearly graded deactivation of the human default-mode network by a visual detection task. NeuroImage, 41:100–112.

53.

Sperling

, Aisen

, Beckett

, et al. 2011. Toward defining the preclinical stages of Alzheimer's disease: recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimers Dement, 7:280–292.

54.

Sporns

. 2011. The human connectome: a complex network. Ann N Y Acad Sci, 1224:109–125.

55.

Stam

, Jones

, Manshanden

, et al. 2006. Magnetoencephalographic evaluation of resting-state functional connectivity in Alzheimer's disease. NeuroImage, 32:1335–1344.

56.

Taulu

, Simola

. 2006. Spatiotemporal signal space separation method for rejecting nearby interference in MEG measurements. Phys Med Biol, 51:1759–1768.

57.

Thomas Yeo

, Krienen

, Sepulcre

, et al. 2011. The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J Neurophysiol, 106:1125–1165.

58.

Uhlhaas

, Singer

. 2006. Neural synchrony in brain disorders: relevance for cognitive dysfunctions and pathophysiology. Neuron, 52:155–168.

59.

Van Veen

, Van Drongelen

, Yuchtman

, et al. 1997. Localization of brain electrical activity via linearly constrained minimum variance spatial filtering. IEEE Trans Biomed Eng, 44:867–880.

60.

Vega

, Hohman

, Pryweller

, et al. 2015. Resting-state functional connectivity in individuals with down syndrome and williams syndrome compared with typically developing controls. Brain Connect, 5:461–475.

61.

Vogel

, Iturria-Medina

, Strandberg

, et al. 2020. Spread of pathological tau proteins through communicating neurons in human Alzheimer's disease. Nat Commun, 11:2612.

62.

Wilcock

, Griffin

WST

. 2013. Down's syndrome, neuroinflammation, and Alzheimer neuropathogenesis. J Neuroinflammation, 10:84.

63.

Wilson

, Vatansever

, Annus

, et al. 2019. Differential effects of Down's syndrome and Alzheimer's neuropathology on default mode connectivity. Hum Brain Mapp, 40:4551–4563.

64.

, Engels

MMA

, Hillebrand

, et al. 2017. Selective impairment of hippocampus and posterior hub areas in Alzheimer's disease: an MEG-based multiplex network study. Brain, 140:1466–1485.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.08 MB

0.21 MB

0.02 MB