Neurologists

Three parts of your work where AI is already doing real lifting, and what stays yours.

• Reflex hammer1860–1929

  Reflex hammer, tuning fork, and clinical observation (pre-EEG era)

  For the first six decades of organized neurology, the neurologist's diagnostic tools were entirely physical: the patella reflex hammer (the rubber hammer arrived in clinical use in the 1880s), the tuning fork for vibration sense and hearing, the ophthalmoscope for fundal examination, the pinwheel for sensory mapping, and above all the trained eye watching how a patient walked, held posture, moved the eyes, and used the hands. Charcot's method had been built on careful observation and correlation with autopsy findings. American neurologists of the Hammond-Mitchell generation applied the same discipline. The neurological examination was not a preliminary to imaging; it was the entire investigation. Lesion localization was a cognitive act performed at the bedside, drawing on detailed anatomical knowledge of the nervous system. The neurologist who could localize a lesion to the posterior inferior cerebellar artery territory from watching a patient walk across a room was doing something genuinely impressive, and it made neurology a specialty that attracted physicians who valued precision thinking over procedural volume.

• EEG1929–1973

  EEG (Hans Berger 1929; US clinical adoption 1940s-1950s)

  Hans Berger recorded the first human electroencephalogram in 1924 and published his findings in 1929. The discovery was confirmed at Harvard by Hallowell Davis in 1934, and the first US clinical EEG laboratories opened in the late 1930s and 1940s. By 1947, the American EEG Society (later the American Clinical Neurophysiology Society) had been founded. In the 1950s, EEG was filtering from academic medical centers into private practices. For the neurologist, EEG was the first technology that gave any functional view of the brain in a living patient. It was transformative specifically for epilepsy: EEG allowed classification of seizure types, identification of focal epileptogenic zones, and monitoring of treatment response in ways the clinical examination alone never could. It also opened a window into encephalopathies, sleep disorders, and brain death determination. The Grass Company's reliable, well-supported EEG machines drove a rapid expansion of clinical knowledge in the 1950s.

  Effect on the work

  EEG created a new procedural competency within neurology and drove the development of clinical neurophysiology as a subspecialty. It did not reduce neurologist headcount but expanded the scope of what neurologists could diagnose and the settings in which they could add value, particularly in epilepsy monitoring units and ICUs.

• CT scanning1973–1984

  CT scanning (first US clinical installations 1973: Mayo Clinic and MGH)

  The first CT brain scan was performed at Atkinson Morley's Hospital in London in 1971. The first US clinical installations went to Mayo Clinic and Massachusetts General Hospital in the summer of 1973. By 1980, millions of CT examinations had been performed in the United States. CT scanning ended the era of purely clinical neurological diagnosis: pneumoencephalograms, ventriculograms, nuclear brain scans, exploratory burr holes, and craniotomies performed to diagnose rather than treat became unnecessary or rare. A neurologist could now confirm or refute a clinical localization hypothesis with a scan result available within hours. The technology did not make the neurologist redundant; it made the neurologist more useful. A confident anatomical localization was now testable, and a neurologist who ordered appropriately was demonstrating a high positive predictive value against a CT ground truth. The specialty's hospital standing improved substantially in the 1970s-1980s as a result.

  Effect on the work

  CT scanning drove an increase in demand for neurological consultations. Hospitals investing in CT scanners needed neurologists on staff who could interpret neurological findings in the imaging context. The 1970s-1980s were a period of specialty expansion, with neurology residency programs growing and subspecialties (neuroradiology, neurocritical care, vascular neurology) beginning to differentiate.

• Clinical MRI1984–1996

  Clinical MRI (first scanners 1977-1980; widespread clinical adoption 1984-1990)

  The first clinical MRI scanners appeared in the early 1980s; by the mid-1980s, MRI of the brain and spine had joined CT as a standard clinical tool, and by 1990 it had largely superseded CT for most non-emergency neurological investigations. MRI's superior soft tissue contrast made it the definitive tool for multiple sclerosis (visible white matter lesions for the first time), posterior fossa lesions (poorly visualized by CT due to bone artifact), spinal cord pathology, and early ischemic stroke. For neurologists, MRI meant that the clinical and anatomical hypothesis reached at the bedside could now be tested with extraordinary precision. Lesions that were clinically suspected but CT-invisible became MRI-visible. MS, which had been diagnosed on clinical criteria alone for decades, could now be confirmed radiographically. This transformed neurology into a precision diagnostic specialty rather than a predominantly syndromic one.

  Effect on the work

  MRI drove another round of demand expansion for neurological consultations. The McDonald criteria for MS (first formulated in 2001, revising earlier 1983 Poser criteria) were built entirely around MRI lesion patterns. The ability to diagnose and monitor MS, brain tumors, vascular malformations, and spinal pathology with MRI made the neurologist's role in multi-disciplinary teams more central.

• tPA stroke thrombolysis1996–2016

  tPA stroke thrombolysis (FDA approved June 1996) and evidence-based neurology protocols

  On June 18, 1996, the FDA approved tissue plasminogen activator (tPA, alteplase) for acute ischemic stroke based on the landmark 1995 NINDS trial. The approval transformed stroke from a condition neurologists watched and managed conservatively into a neurological emergency where speed of diagnosis and treatment directly determined outcome. The "time is brain" framework (more than a million neurons die per minute in a large vessel occlusion) made the neurologist's ability to rapidly assess, image, and treat an acute stroke the central value proposition of hospital-based vascular neurology. This era also saw the formalization of stroke code protocols, the development of comprehensive stroke centers, and the growth of vascular neurology as a certified subspecialty (ABPN vascular neurology certification introduced 2005). The tPA window (later extended to several hours by ECASS III in 2008) made the neurologist's role on-call and time-sensitive in a way that no prior neurological treatment had been.

  Effect on the work

  tPA approval drove a sustained increase in demand for vascular neurologists and stroke-specialized practitioners. Hospitals seeking Joint Commission Primary Stroke Center certification (program launched 2003) needed dedicated neurological coverage. The number of trained vascular neurologists grew substantially in the 2000s and 2010s.

• EHR adoption2012–2019

  EHR adoption (Meaningful Use 2011-2015) and the documentation burden era

  The HITECH Act (2009) funded EHR adoption across American hospitals through the Meaningful Use program. By 2015, most of US hospitals had adopted certified EHR systems. For neurologists, EHR adoption had a deeply ambivalent effect. The upside was instant access to imaging results, lab values, prior notes, and medication histories across a care episode. The downside was a documentation burden so severe that physicians began reporting that they spent more time in the EHR than with patients. The neurology consult note is among the most information-dense in hospital medicine: it requires synthesis of a complex symptom history, a detailed neurological exam with explicit findings across all cranial nerves and motor-sensory-cerebellar-gait domains, interpretation of imaging, a differential diagnosis, and a management plan. EHR point-and-click templates were poorly adapted to the neurologist's way of thinking, and many neurologists spent a few hours per consult on documentation. Burnout rates in neurology rose substantially during this era.

  Effect on the work

  AAN surveys from 2012-2018 documented rising burnout rates among neurologists, attributed substantially to EHR documentation burden. The documentation overhead reduced effective patient-facing capacity per neurologist, worsening the workforce shortage without reducing the number of employed neurologists.

• AI neurovascular triage2016–2023

  AI neurovascular triage (Viz.ai 2016-present; RapidAI, Aidoc, Avicenna CINA)

  Viz.ai received its first FDA 510(k) clearance for AI-assisted large vessel occlusion detection in 2018 and was deployed at over 1,200 hospitals globally by 2026. RapidAI (RAPID), Brainomix e-Stroke, and Aidoc built comparable platforms. These tools receive CT and CTA data directly from the PACS, run AI inference within minutes, and send a mobile push notification to the on-call vascular neurologist with an LVO detection, ASPECTS score, and ischemic core/penumbra mismatch map. The clinical effect was compressing the call-to-neurologist-awareness time from many minutes (a radiologist reading the scan and calling) to several minutes. Viz.ai's published real-world data showed a median workflow savings compared to standard radiologist read. For a specialty where more than a million neurons die per minute in a stroke, this was not an incremental improvement; it was a structural shift in what it was possible to accomplish in a given patient's time window.

  Effect on the work

  AI neurovascular triage tools extended one neurologist's effective reach across hub-and-spoke hospital networks without requiring on-site presence. A vascular neurologist at a comprehensive stroke center could receive AI alerts from multiple spoke hospitals simultaneously, review imaging on a mobile device, and direct care remotely. The tools did not reduce demand for neurologists; they expanded what a fixed number of neurologists could cover.

• Ambient AI documentation2023–present

  Ambient AI documentation and clinical decision support (Dragon Copilot, OpenEvidence, Glass Health)

  Dragon Copilot (Microsoft-Nuance) began widespread deployment at US health systems in 2023-2025, with neurology-specific templates for consult notes, discharge summaries, and follow-up visits. A 2025 JAMA Network Open randomized controlled trial confirmed a 62% reduction in after-hours EHR documentation time for physicians using ambient AI scribes. A parallel Scientific Reports 2025 study specifically validated that AI-generated neurology consultation summaries improved efficiency and reduced documentation burden in the emergency department. OpenEvidence, used by over 757,000 verified US physicians daily by 2026, provides instant AAN guideline lookup at the bedside for complex protocols (tPA eligibility thresholds, anti-amyloid therapy ARIA monitoring, AED selection). Glass Health structures differentials for complex presentations that neurology is frequently called to evaluate. Together, these tools are beginning to recapture the documentation and cognitive-overhead burden that grew with EHR adoption, returning time to direct patient care.

  Effect on the work

  Ambient documentation tools do not reduce the need for neurologists; they increase the effective patient-facing capacity of each neurologist. The 62% reduction in after-hours documentation time represents hours per week returned to clinical work, sleep, or sustainable practice. For a specialty with a structural workforce shortage, this augmentation effect is more materially valuable than the displacement concern.