This dissertation models the relationship between symptoms of heart rhythm fluctuations and cardiac measurements in order to better identify the probabilities of either a primarily organic or psychosomatic cause, and to better understand cognition of the internal body. The medical system needs to distinguish patients with actual cardiac problems from those who are misperceiving benign heart rhythms due to psychosomatic conditions. Cognitive neuroscience needs models showing how the brain processes sensations of palpitations. Psychologists and philosophers want data and analyses that address longstanding controversies about the validity of introspective methods. I therefore undertake a series of measurements to model how well patient descriptions of heartbeat fluctuations correspond to cardiac arrhythmias. First, I employ a formula for Bayesian inference and an initial probability for disease. The presence of particular phrases in symptom reports is shown to modify the probability that a patient has a clinically significant heart rhythm disorder. A second measure of body knowledge accuracy uses a corpus of one hundred symptom reports to estimate the positive predictive value for arrhythmias contained in language about palpitations. This produces a metric representing average predictivity for cardiac arrhythmias in a population. A third effort investigates the percentage of patients with palpitations report actually diagnosed with arrhythmias by examining data from a series of studies. The major finding suggests that phenomenological reports about heartbeats are as or are more predictive of clinically significant arrhythmias than non-introspection-based data sources. This calculation can help clinicians who must diagnose an organic or psychosomatic etiology. Secondly, examining a corpus of reports for how well they predict the presence of cardiac rhythm disorders yielded a mean positive predictive value of 0.491. Thirdly, I reviewed studies of palpitations reporters, half of which showed between 15% and 26% of patients had significant or serious arrhythmias. In addition, evidence is presented that psychosomatic-based palpitation reports are likely due to cognitive filtering and processing of cardiac afferents by brainstem, thalamic, and cortical neurons. A framework is proposed to model these results, integrating neurophysiological, cognitive, and clinical levels of explanation. Strategies for developing therapies for patients suffering from identifiably psychosomatic-based palpitations are outlined. / text