Fig 1.
Data overview: Cross-sectional age distribution of the ‘10kInOneDay’ dataset used in this study with additional details of the dataset.
We consider a total of N = 2018 participants, which for some analyses we split to further categorize the ‘Young’ (<30 years), ‘Middle’ (≥30 & <60 years), and ‘Old’ (≥60 years) ages. Each participant is healthy. The dataset is multi-sited across 6 different research groups.
Fig 2.
a) The excitatory time series for a single structural connectome after perturbing the left hemisphere insula subregion 1 brain region (belonging to the fronto-parietal cognitive system) with strength P = 1.15 for the entire duration. Top half shows the time series for the individually stimulated brain region and the bottom half shows the responding brain regions’ time series. b) Single snapshot in time of the phases (Eq (8)) represented on the unit circle for a cognitive system (CS) pair, , which are used to determine the amplitude of the complex Kuramoto order parameter (Eq (9)). Specifically, the brain regions that belong to the cingulo-opercular (CO) and motor-sensory (MS) cognitive system are shown, with CO in light pink and MS in dark pink. The length of the black vector arrow represents the amplitude of the complex Kuramoto order parameter,
, measuring the synchronization with respective to the CO and MS cognitive systems. This example gives a value of
for this snapshot in time. c) Time series of the amplitude of the complex Kuramoto order parameter,
. The horizontal line in red,
, represents the mean over the time window,
s, corresponding to the synchrony (Eq (11)) between the CO and MS cognitive systems. d) Matrix of the synchrony values for all pairwise combinations of the nine cognitive systems. The specific example for the synchrony between CO and MS is outlined with a black square. The remaining cognitive systems are labeled attention (Att), auditory (Aud), fronto-parietal (FP), default mode (DM), subcortical (SC), ventral-temporal (VT), and visual (V) for the synchrony matrix.
Fig 3.
Classification of the emergent synchronization patterns when stimulating a single brain region for three different cases.
The left column corresponds to stimulating the left hemisphere banks of the superior temporal sulcus (BANKSSTS), the middle columns is for the left hemisphere pallidum, and the right column is for the hemisphere insula subregion 1, all for the same structural connectome. Panels a), b), and c) show their respective pair-wise synchrony matrices. Panels d), e), and f) illustrate the binarization of the synchrony matrices using a threshold of , with
representing a synchronous element (red) and an asynchronous element (blue), otherwise. Panels g), h), and i) show the community assignments that occur when applying the Louvain algorithm. Cognitive systems belonging to the same community have the same colour and are labeled with the same number. Panels j), k), and l) show the transformation of the cognitive systems’ community assignments into synchronization patterns, where blue colors indicate cognitive systems in the asynchronous group and red colors indicate cognitive systems in the synchronous group. The emergent synchronization patterns and, hence, the stimulated brain region are then classified as asynchronous (containing only an asynchronous group), synchronous (containing only a synchronous group), and chimera (containing both synchronous and asynchronous groups). The cognitive systems are labeled attention (Att), auditory (Aud), cingulo-opercular (CO), fronto-parietal (FP), default mode (DM), motor-sensory (MS), subcortical (SC), ventral-temporal (VT), and visual (V).
Fig 4.
Proportions and age-related differences in the proportions of the three synchronization pattern classes (asynchronous (async), sychronous (sync) and chimera, see Fig 3) associated with stimulated brain regions (coarse-grained atlas).
The first column (panels a), e), and i)) shows the average classification proportion with respect to the cross-sectional lifespan. The second column (panels b), f), and j)) shows the change in classification proportion from young age (<30 years) to middle age (≥30 & <60 years) and the third column (panels c), g), and k)) the change in classification proportion from middle age to old age (≥60 years), whereas the last column (d), h), and l)) show the net change from young age to old age. The first row (panels a), b), and c), d)) corresponds to the asynchronous classification proportions, the second row (panels e), f), g), and h)) corresponds to chimera state classification proportions, and the third row (panels i), j), k), and l)) corresponds to synchronous classification proportions. Each column contains three views for the brain: Lateral, medial, and subcortical, respectively from left to right. The lateral and medial view portray the right and left hemispheres, respectively from up to down, and the subcortical shows the front. This figure was drawn using the open-sourced ‘ggseg‘ package in R [77].
Fig 5.
The classification proportions across cross-sectional age over five select cognitive systems: a) auditory-left (Aud-L; dark blue), b) auditory-right (Aud-R; dark purple), c) ventral temporal (magenta), d) default mode (pink), and e) attention (orange).
Blue circles represent the asynchronous classification proportions, yellow squares represent the chimera state classification proportions, and red triangles represent the synchronous classification proportions. The first column shows the brain region labels that belong to the respective cognitive system for each cognitive system, the second column shows where the brain regions are located spatially, and the third column shows how the classification proportions change with cross-sectional age. Shaded regions are 95% confidence intervals (CI) across individuals. Here and throughout, these are calculated as , where
is the mean of the classification proportion for an age group, t is the critical t-value corresponding to 95% confidence intervals, σ is the standard deviation of the classification proportion for an age group, and n is the number of individuals for an age group. The maps of brain regions were drawn using the open-sourced ‘ggseg‘ package in R [77].
Fig 6.
The variation in classification proportions with strength of the stimulated brain region belonging to the respective cognitive system in a given age group including the strength distribution (curve with shaded area) along with its quartile as a box plot below.
Blue circles represent the asynchronous (async) class, yellow squares represent the chimera class, and red triangles represent the synchronous (sync) class. Regions shaded in blue, yellow, and red are dominated by the asynchronous, chimera, and synchronous classification, respectively. Each column represents a different cognitive system a) for Auditory-Left (dark blue, b) for Auditory-Right (dark purple), c) for Ventral-Temporal (magenta), d) for Default Mode (pink), and e) for Attention (orange). Each row corresponds to a different age group, from young (<30 years) to middle-aged (≥30 % <60 years) to old age (≥60 years).
Fig 7.
The variation in the dominating classification proportions with age and strength of the stimulated brain region belonging to the respective cognitive system including the classification proportions and strength distributions for a) the auditory-left, b) auditory-right, c) ventral-temporal, d) default mode, and e) attention cognitive systems.
Shaded regions represent the dominating classification at that strength range (bottom x-axis) with blue, yellow, and red representing asynchronous, chimera, and synchronous classes, respectively. Symbols correspond to the classification proportions as in Fig 5 (blue circles: asynchronous (async) class, yellow squares: chimera class, red triangles: synchronous (sync) class, and top x-axis label). The strength distributions are shown (grey) along with their quartiles as a box plot.
Fig 8.
Proportion of emergent synchronization patterns along cross-sectional age ( participants for each age group) for five different cognitive systems.
The first column (panels a), c), e), g), i)) shows the proportion of prevalent patterns (≥5% in at least one age window). As in Fig 3, a synchronization pattern with a red (blue) block represents the synchronous (asynchronous) class. The second column (panels b), d), f), h), j)) shows the proportion of patterns ordered by rank. Tiles containing an ‘X’ occur with ≥5% prevalence and ‘O’ occur with ≥3% prevalence. Each row represents a different cognitive system: auditory-left (panels a-b; dark blue), auditory-right (panels c-d; dark purple), ventral-temporal (panels e-f; dark magenta), default mode (panels g-h; pink), and attention (panels i-j; orange).
Fig 9.
a) Variations in similarity of synchronization patterns, as defined in Eq (12), with cross-sectional age across the Auditory-Left (Aud-L, dark blue), Auditory-Right (Aud-R, dark purple), Ventral-Temporal (VT, magenta), Default Mode (DM, pink), and Attention (Att, orange) cognitive systems.
b) Same as a) but for only the synchronization patterns classified as chimera.
Fig 10.
Region similarity and individual similarity across the three age groups: a) young (<30 years, b) middle-aged (≥30 & <60 years), and c) old age (≥60 years) for all cognitive systems: attention (Att), auditory (Aud (left and right)), cingulo-opercular (CO), fronto-parietal (FP), default mode (DM); motor-sensory (MS), subcortical (SC), ventral-temporal (VT), and visual (V).
Error bars show 95% confidence intervals. Note that the error bars in the x-direction are typically smaller than the symbol sizes.