A brain-computer interface (BCI) is a system that allows communication between a subject's brain and a computer, without requiring any physical movement. The high-level process for any BCI system is signal acquisition, signal processing and providing output through a physical device. Signal acquisition is done by sensors with multiple channels using invasive or non-invasive methods. The collected data usually contains more than 256 unique channels across the standard 32 sensors. In addition to this, the data is often sampled at a rate of 256Hz for long periods of time, resulting in massive datasets. The critical next step is to preprocess these signals to understand the patterns and changes in the various types of brain waves. However, the intrinsic neurophysiological changes in the brain pose a challenge to the processing of these signals. Identifying the key sensor channels becomes vital not only for the above reason but also due to the large size of the data points captured. This paper is a study on identification of optimal channels for the detection of an emotion such as Fear, one of the primary emotions frequently observed during multiple studies. By detecting channels with highly prevalent features, this study aims to reduce the size of data. The data so optimized can be output to speed up complex classification/prediction algorithms.