This paper proposes a self-reset pulse frequency modulation (PFM) digital pixel sensor (DPS) with in-pixel variable reference voltage for optical brain imaging systems. The sensor demonstrates a wide dynamic range and very low power consumption that can detect small signals of brain activity in brain. The high dynamic range, high SNR (signal-to-noise ratio), high speed and low power consumption image sensor are suitable for optical brain imaging systems. Since the comparator part consumes high power inside pixel, sub-threshold, self-biased and bulk-driven techniques are used to achieve both ultra-low-voltage and low power in the PFM DPS. Moreover, High speed (high frame rate) is achieved by image capturing in-parallel for all pixels. The proposed image sensor is post layout simulated in 0.18 µm Complementary Metal Oxide Semiconductor (CMOS) technology with 0.6 V supply voltage, resulting in the dynamic range of 152 dB and the power consumption of 11.25 nW and the fill factor of the proposed sensor is 11%. Hence, this device has significant potential to be used for brain signal detection in pre-clinical and clinical studies, cognitive process, diagnose diseases in exploring brain structure and function. [ABSTRACT FROM AUTHOR]