At the NIMH, we have continued to collaborate with Professors Jules Angst and Wulf Rossler to analyze this unique longitudinal prospective study of a cohort of young adults who have been followed for 25 years since the age of 18. During the past year we focused on assessing the validity of the definitions of headache subtypes, their longitudinal stability, and patterns of co-occurrence with mental disorders. The results confirm those of prior studies that document the high prevalence and impact of migraine and other headache subtypes in the general population. The substantial longitudinal overlap across headache subtypes suggests that migraine may not be a homogenous entity and that etiologic studies, particularly treatment and genetic studies that are based on the distinction between life time diagnoses of migraine with and without aura based on retrospective classification may not capture the true nature of headache in the general population. Finally, the gradient of severity across headache subtypes for treatment, medication use, impairment, severity and chronicity suggest that the various headache subtypes may be better characterized on a spectrum rather than as distinct clinical entities. Our analyses of the longitudinal patterns of co-occurrence of migraine with mood and anxiety disorders reveals that migraine is specifically associated with social phobia, panic disorder and bipolar disorder. These specific findings confirm the results in a cross-sectional sample of the U.S. National Comorbidity Survey-Replication(NCS-R). These findings have important implications for the health policy, classification and etiologic studies of migraine and other headaches. First, our findings confirm the abundant evidence from prior cross sectional and retrospective and short-term prospective studies of population-based samples regarding the high prevalence rates and dramatic impact of migraine on subjective distress and work impairment, as well as the high frequency of contact with the health care system. Second, the substantial longitudinal overlap across headache subtypes also suggests that migraine may not be a homogenous entity and that etiologic studies, particularly treatment and genetic studies that are based on the distinction between life time diagnoses migraine with and without aura based on retrospective classification may not capture the true nature of headache in the general population. With the dramatic fluctuations in expression across headache subtypes over time, research requiring a lifetime diagnosis might be better served by classification of the person across time rather than by acute episodes that may change across the life span. Finally, the findings regarding patterns of comorbidity have important implications for future studies of the biologic pathways underlying these conditions.