In this paper we apply spectral analysis methods to heart rate variability to assess the autonomic nervous system activity in normal subjects and in patients affected by different degrees of diabetic autonomic neuropathy. The current opinion, based on different clinical tests, is that parasympathetic impairment occurs earlier in autonomic dysfunctions. However, the use of power spectrum density analysis based on a single parameter (heart rate) suggests a simultaneous involvement of parasympathetic and sympathetic pathways leading to the conclusion that perhaps the natural history of diabetic autonomic neuropathy should be substantially rewritten.