Human confidence reflects a positive self-perception and balanced autonomic nervous system response. In this paper, we present a three stage approach to detect human confidence level by computing the heart rate from speech signals. First stage involves extraction of breathing patterns from speech using a pre-trained model followed by stage 2, where the heart rate is extracted using Independent Component Analysis (ICA) on the breathing patterns. Finally in stage 3, an analysis of heart rate values indicating human confidence levels is done. To the best of our knowledge, this is the first time ever, with our experiments, empirically it is found that heart rate extracted from speech carries information related to the confidence of the candidate. We observe that, on an average, confident speakers have 10 beats per minute lower heart rate as compared to non confident speakers.