# Coherence Analysis of Cardiovascular Signals for Detecting Early Diabetic Cardiac Autonomic Neuropathy: Insights into Glycemic Control

## TL;DR

This study explores how coherence analysis of cardiovascular signals can detect early signs of diabetic cardiac autonomic neuropathy and assess the impact of insulin treatment.

## Contribution

The study introduces coherence analysis between SBP and RRI as a potential marker for early diabetic cardiac autonomic neuropathy.

## Key findings

- Early diabetic rats showed reduced low- and high-frequency coherence compared to controls.
- Insulin treatment partially restored coherence but did not fully recover it to normal levels.
- Low-frequency power increased at diabetes onset and decreased after insulin therapy.

## Abstract

Background: Cardiac autonomic neuropathy (CAN) is a common yet frequently underdiagnosed complication of diabetes. While our previous study demonstrated the utility of multiscale cross-approximate entropy (MS-CXApEn) in detecting early CAN, the present study further investigates the use of frequency-domain coherence analysis between systolic blood pressure (SBP) and R-R intervals (RRI) and evaluates the effects of insulin treatment on autonomic function in diabetic rats. Methods: At the onset of diabetes induced by streptozotocin (STZ), rats were assessed for cardiovascular autonomic function both before and after insulin treatment. Spectral and coherence analyses were performed to evaluate baroreflex function and autonomic regulation. Parameters assessed included low-frequency power (LFP) and high-frequency power (HFP) of heart rate variability, coherence between SBP and RRI at low and high-frequency bands (LFCoh and HFCoh), spontaneous and phenylephrine-induced baroreflex sensitivity (BRSspn and BRSphe), HRV components derived from fast Fourier transform, and MS-CXApEn at multiple scales. Results: Compared to normal controls (LFCoh: 0.14 ± 0.07, HFCoh: 0.19 ± 0.06), early diabetic rats exhibited a significant reduction in both LFCoh (0.08 ± 0.04, p < 0.05) and HFCoh (0.16 ± 0.10, p > 0.05), indicating impaired autonomic modulation. Insulin treatment led to a recovery of LFCoh (0.11 ± 0.04) and HFCoh (0.24 ± 0.12), though differences remained statistically insignificant (p > 0.05 vs. normal). Additionally, low-frequency LFP increased at the onset of diabetes and decreased after insulin therapy in most rats significantly, while MS-CXApEn at all scale levels increased in the early diabetic rats, and MS-CXApEnlarge declined following hyperglycemia correction. The BRSspn and BRSphe showed no consistent trend. Conclusions: Coherence analysis provides valuable insights into autonomic dysfunction in early diabetes. The significant reduction in LFCoh in early diabetes supports its role as a potential marker for CAN. Although insulin treatment partially improved coherence, the lack of full recovery suggests persistent autonomic impairment despite glycemic correction. These findings underscore the importance of early detection and long-term management strategies for diabetic CAN.