Cognitive and motor functions of the human brain often display non-linear dynamics. Fractal dimension such as Higuchi’s fractal dimension (HFD) has been used as an effective method to probe these nonlinear dynamics and has been used to characterise different diseases. However, there is inconsistency in how HFD varies with age and disease in literature, which could be due to different computational techniques or frequency ranges used in these studies. To resolve this disparity, we investigated the frequency-dependant change of HFD with age for 64-channel electroencephalogram (EEG) data recorded from 217 elderly human population, aged 50-88 years, for baseline eyes open state and during the presentation of the achromatic grating stimulus at high sampling frequency (2.5 kHz). In the baseline eyes open state, with age, HFD increased at frequencies upto 200 Hz and showed an opposite trend at higher frequencies. Interestingly, this change in HFD was opposite to the age-related change in 1/f slope of the power spectral density (PSD). Moreover, the stimulus-related change in HFD was anti-correlated with the change in oscillatory power in alpha (8-12 Hz) and gamma (20-66 Hz) ranges. Therefore, HFD effectively captures changes in the oscillatory signatures such as gamma oscillations as well as the 1/f slope of the power spectrum. Since they both are known signatures of excitation-inhibition (E-I) balance in neural circuits, HFD could provide an effective measure of E-I balance which is altered with several neurological disorders and could, therefore, provide a useful biomarker of brain function in health and disease.