Detection of Early Diabetes Using Non-Invasive Multi-Electrode Bioimpedance Recording
Divij Tikoo
11/08/2026
Diabetes mellitus is among the fastest-growing chronic diseases worldwide, affecting hundreds of millions of people and placing an increasing burden on global healthcare systems. Early diagnosis remains essential for preventing long-term complications, yet conventional diagnostic techniques such as fasting blood glucose testing, oral glucose tolerance tests, and glycated haemoglobin (HbA1c) measurements rely on invasive blood sampling and may not be suitable for large-scale or continuous screening. Consequently, there is growing interest in developing non-invasive, affordable, and accessible diagnostic technologies.
This study investigates the feasibility of multi-frequency tetrapolar bioimpedance spectroscopy as a non-invasive approach for distinguishing between healthy and diabetic individuals. Impedance magnitude, phase angle, and reactance were measured from thirty participants using an AD5933-based bioimpedance analyzer with a four-electrode (tetrapolar) configuration. Measurements were acquired at 10 kHz, 50 kHz, 100 kHz, and 500 kHz to evaluate frequency-dependent electrical behaviour associated with diabetes.
Experimental observations demonstrated consistently higher impedance values in participants with diabetes, particularly at lower frequencies where extracellular conduction predominates. Statistical analysis confirmed significant differences between the healthy and diabetic groups across multiple frequencies, while phase angle and reactance measurements exhibited trends consistent with altered membrane capacitance and tissue dielectric properties. Based on these findings, two exploratory analytical frameworks are proposed: the Frequency-Dependent Diabetic Electrical Signature (FD-DES), describing the characteristic impedance profile observed across frequencies, and the Diabetes Impedance Separation Index (DISI), a quantitative metric for expressing impedance separation between healthy and diabetic populations.
Although the present investigation represents a pilot study and further large-scale clinical validation is required, the results suggest that multi-frequency bioimpedance spectroscopy may provide a promising foundation for future non-invasive diabetes screening systems. The integration of bioimpedance sensing with wearable electronics and artificial intelligence may ultimately enable continuous, low-cost metabolic health monitoring in both clinical and community settings.