HIV encephalitis remains a prevalent and important neuropathological correlate of dementia in the HAART era. It is probably the only neuropathological variable that has undergone substantial independent verification in multiple autopsy cohorts. HIV encephalitis has provided the field with a unifying concept to pursue the pathophysiology of dementia. Drawbacks remain that the diagnosis is an overly diverse constellation of anomalies in both gray and white matter, and correlation with dementia remains too vague. The overall incidence of HIV-associated dementia in the HAART era declined by about half, but autopsy cohorts still exhibit a substantial prevalence of HIV encephalitis, probably because they are enriched with people with end-stage disease. The intensity of HIV encephalitis might have undergone a decrease in the HAART era, but HIV encephalitis remains essentially an "all or none" diagnosis. The criteria for a diagnosis of HIV encephalitis remain essentially intact but variants might have appeared in the HAART era. Examples include an increase in the number of cases with predominantly white matter necrosis ("leukoencephalopathy variant"), cases with only minimal or pathological change ("functional variant"), cases that have scars that lack evidence of HIV-1 after treatment with HAART ("burnt out variant"), cases with age-associated degeneration ("elderly variant") and cases in which immune reconstitution inflammatory syndrome has played a role ("inflammatory variant"). The partial disassociation between clinical dementia and HIV encephalitis remains a longstanding paradox of the neuro-AIDS field. It suggests that more than one underlying pathophysiological mechanism is present, and that co-morbid factors may be important. Progress in understanding the cliniconeuropathological correlation should advance because brain specimens collected under standardized conditions are available from the National NeuroAIDS Tissue Consortium. The incidences of opportunistic infection and tumors in the CNS underwent decreases after HAART. HAART prevents these processes from developing, but initiation of therapy when there is an ongoing pathological process can provoke immune reconstitution (immune reconstitution inflammatory syndrome) and an intense pathological response in the CNS that is potentially fatal. HIV/AIDS might exacerbate diseases of brain aging, and aging could worsen HIV-1-associated neurocognitive disability. The neurobiological interaction between persistent HIV-1 infection and brain aging remains unclear. Brain aging is universal to all populations, so neuropathological study in people with HIV/AIDS must control for chronological age. In the peripheral nervous system painful H-DSP is the most prevalent neurological problem of HIV/AIDS in the HAART era. The distal dying-back small fiber type of neuropathy still produces a disabling neuropathic pain. Measuring the loss of distal nerve fibers in the epidermis of a skin biopsy is a new diagnostic approach that is potentially useful to track disease progression and/or nerve regeneration experimentally. Even though the pathology occurs in the distal terminals, it remains likely that the driving force is a more proximal dorsal root ganglionopathy that is not characterized neuropathologically, or experimentally. Signs of HIV encephalitis and other opportunistic infection in the spinal cord have declined in the HAART era. The prevalence of HIV-associated vacuolar myelopathy has waned primarily because the lesion was predominantly observed in end-stage of AIDS, which is less prevalent. Gracile tract atrophy remains prevalent, and probably results from both sensory peripheral neuropathy and aging.