When the inner clock fades: Interoceptive decline and consolidation of phase resetting in cortical rhythms by cardiac events underlie healthy lifespan aging

Figures

• 

  Fig. 1.. Overview of the Data Pre-processing and Analysis Pipeline. (a) Simultaneous MEG-ECG recordings undergo a preprocessing pipeline including detrending, bandpass filter and artifact removal using Independent Component Analysis (ICA). R peaks are detected from the electrocardiogram (ECG) waveform using Pan-Tompkin algorithm, serving as a trigger point to compute heartbeat evoked responses (HERs). (b) HERs are computed across all the age-groups (young, middle and old). Additionally, wavelet analysis is employed to investigate the mechanisms underlying HERs generation. (c) Source localization using eLORETA identifies cortical sources underlying HERs. (d) Finally, Granger causality maps casual functional relationships among cortical sources of HERs.

• 

  Fig. 2.. Cardiovascular measures across age. (a) Methods used to capture cardiac activity includes electrocardiogram (ECG) and sphygmomanometer which measures cardiovascular activities such as electric activity of the cardiac muscles and blood pressure respectively. Cardiovascular parameters include Inter beat interval (IBI), Heart rate variability (HRV), Heart rate (HR), Systolic and Diastolic blood pressure. (b) IBI significantly decreases (r = -0.0994; p = 0.0133). (c) HRV doesn’t change (r = -0.0022; p = 0.9573). (d) HR significantly increases (r = 0.1121; p = 0.0052). (e) Systolic blood pressure significantly increases (r = 0.4370; p < 0.001) and (f) Diastolic blood pressure remains unchanged (r = 0.0734; p = 0.0910) across age.

• 

  Fig. 3.. Heartbeat evoked responses (HERs) across age and gender. (a) Method: Simultaneous MEG-ECG recording: using R peak of the ECG waveform as the time points to epoch neural time series to obtain the HERs. (b) Group-wise analysis of HERs obtained from the time locked MEG activity to R peak of the ECG waveform. Statistical test conducted indicates that the HERs differences across the age groups are significant shown by the asterisks (F(2,619) = 9.48; p < 0.001). (c) Group wise analysis of HERs obtained from time locked activity to the shuffle ECG time series (sanity check). (d) Linear regression conducted between magnitude (obtained from peak to peak subtraction) of HERs and age (r = -0.1963; p < 0.001)

• 

  Fig. 4.. Mechanism of Heartbeat Evoked Responses. Inter trial phase coherence (ITPC) was computed on evoked data time locked to (a-c). R peak (d-f). shuffled R peak of ECG waveform calculated across all the age group. ITPC values within the white boundary are considered to be statistically significant, using critical value at p = 0.05. Time frequency spectrograms calculated on evoked data time locked to (g-i). R peak (j-l). shuffled R peak of ECG waveform (found no significant difference between the age groups; cluster permutation test at p = 0.05). (m) ITPC value (extracted from each subject) in theta band significantly increases across age (r = 0.2236; p < 0.001). (n) No correlation was observed between mean ITPC values and mean power extracted at around 180–320 ms post R peak of ECG waveform for each participant (r = 0.0596; p = 0.1382).

• 

  Fig. 5.. Cortical sources of Heartbeat evoked responses (HERs) across age: Continuous and categorical analysis. The sources represent the cortical areas involved in the processing of heartbeat which includes frontal-temporal regions across age. Panel (a-c) represent sources derived from categorical analysis, while panels (d-g) show results from continuous analysis. The statistical threshold was set at 95th percentile and the source powers of grid points to surpassing this threshold were considered as significant sources of activation. All regions were approximated to the nearest Broadmann areas of the human brain. Regression between source-level HERs and age for different regions are (h-r) Frontal, (s-u) Temporal and (v-w) Cingulum and insula.

• 

  Fig. 6.. Cortico-cortical and cortico-cardiac information flow across lifespan. (a) Workflow used for the calculation of Granger Causality (GC) to evaluate information flow from brain-brain and brain-body interactions. (b) Visualization of brain nodes on a glass brain template, used for subsequent cortico-cortical connectivity calculation. (c) GC heatmaps showing connectivity between brain nodes; red indicated high GC value while blue indicates lower GC values. Black boxes highlight statistically significant values after permutation testing followed by Bonferroni correction. (d) Connectivity maps displaying significant connection between brain nodes (red line: strong connectivity; blue line: weak connectivity). (e) Age-related trends in connectivity between and within OFC and TP; connectivity from R mid OFC (r = 0.1846; p < 0.0001) and R inf OFC (r = 0.0891; p = 0.0269) increases to R med OFC across age; the connectivity from L mid OFC (r = -0.1037; p = 0.0099) to R mid TP decreases across age. (f) Cortico-cardiac connectivity: connection from the heart to brain increases across lifespan mostly, while connectivity from brain to heart decreases across age.