Frontotemporal dementia (FTD), a common type of young-onset dementia, has long been described in broad clinical umbrellas: behavioral variant FTD, characterized by prominent changes in personality and social conduct, and primary progressive aphasia, defined by language impairment. These clinical syndromes are coupled to gross anatomical patterns: frontal-predominant atrophy linked to behavioral change, left hemisphere involvement to language dysfunction.
Yet one anatomical pattern did not neatly fit these categories: predominant degeneration of the right anterior temporal lobe. Unlike left temporal disease, it does not present as a classical language syndrome. Unlike frontal-predominant disease, it is not disinhibition, impulsivity, or executive dysfunction. For decades, this phenotype remained clinically diffuse, folded into broader behavioral categories, interpreted through psychiatric lenses, or described only partially.
When we launched the International Working Group on Right Temporal Variant Frontotemporal Dementia (IWG-rtvFTD) in 2020, we were driven by a simple question: How do clinicians actually recognize this syndrome at the bedside? This might sound like a historical curiosity in 2026, when the biomedical landscape is defined by fluid and blood biomarkers, multimodal neuroimaging, and advanced computational models. Yet our field faces a paradox: technological innovation increasingly outpaces our ability to describe patients clearly in the first place.
The syndrome we focus on, FTD with predominant right anterior temporal lobe neurodegeneration, has been described for decades in case reports, cohort studies, and conference hall conversations. The reason it was not more clearly delineated earlier may lie in the complexity of its behavioral and cognitive changes. These patients show a selective erosion of socioemotional semantic knowledge: a diminished understanding of people, social concepts, and emotional meaning. Their social responses shift: not necessarily becoming impulsive, but misaligned. Their motivations alter: long-held values, priorities, and personal preferences reorganize. Such features are subtle, layered, and easily subsumed under broader diagnostic umbrellas. When we think only in gross categories, the phenotype always “fits somewhere.” When we examine it closely, it reveals a distinct and coherent pattern.
Families often struggle to explain “what changed”, and clinicians grapple with whether the core problem is psychiatric, neurological, or perhaps merely a temporary stress reaction to life circumstances. Despite this, the literature has remained fragmented. The same phenomena were labeled as “empathy loss,” “emotional flatness,” “coldness,” “emotional semantic deficits,” or simply “not quite dementia.”
What was missing wasn’t data—it was a coherent neuroscientific framework.
A Clinician’s Problem, A Clinician’s Solution
Our effort began informally: clinicians sharing puzzling cases across continents. What initially looked like anecdotal storytelling quickly became evidence generation. Cases became patterns. Patterns became constructs. Constructs became preliminary data. And these data, enriched by evidence from the literature alongside expert opinions, are what eventually evolved into the consensus paper we just published in Communications Medicine.
One of the underlying messages of our work is that medicine still progresses through careful clinical observation. A patient, not an algorithm, is often the strongest source of biological insight. In this sense, the project is strangely “old-fashioned”: a reminder that a single richly interrogated case can still advance science. Our round-table meetings sometimes spent an hour on one patient. We debated whether the patient’s personality change stemmed from the visual facial recognition deficit, whether the response was blunted or simply misinterpreted, or whether the patient’s ability to conceptualize socio-emotional information was absent. Many of these nuances do not appear in structured datasets. They live in clinics, in family interviews, and in multidisciplinary discussions that only clinicians know how to have.
Why Now? Why This Paper?
The field is at a turning point. Biomarkers, blood-based assays, and machine learning are finally ready to give us mechanistic clarity, but only if we tell them what to look for. Without culturally sensitive, precise clinical definitions, a model cannot predict what we cannot name.
Our paper is the first attempt to provide a neural mechanistic vocabulary for unique symptoms specific to the right anterior temporal lobe neurodegeneration, transforming vague behavioral descriptions into constructs that can be quantified, modeled, and eventually biologically explained. Most strikingly, it demonstrates that the field can move forward by synthesizing knowledge creatively and meticulously from clinicians and families.
The Working Group as a Scientific Instrument
One unanticipated outcome of this project has been the realization that the working group itself functions as an epistemic tool. Behavioral neurologists, psychiatrists, neuropsychologists, and speech-language pathologists rarely sit together and dissect the same cases. When they do, the discipline advances.
This collaborative format also sheds light on why this field has historically lagged: unlike other areas of medicine, behavioral neurology has rarely built large international consortia dedicated to harmonizing clinical constructs. This paper represents the first attempt in our subfield to harmonize nomenclature, data, and clinical constructs across cultures in a neuroscientifically informed way. The working group comprises more than 100 clinicians from over 50 centers across North and South America, Europe, Asia, and Australia, not only by sharing clinical, neuroimaging, biomarker, or genetic data, but also by sharing clinical reasoning.
The paper reminds us that clinical phenomenology is not a relic, it is the foundation on which biological and computational discovery rests.
We often say in neurology that “patients are our teachers.” This study proves that the statement is not sentimental, but methodological. The field moves when clinicians talk to each other, argue over cases, and remain humble before the complexity of the human mind.
In an era of biomarkers, pipelines, and AI prediction models, this work’s contribution is refreshingly simple: before we measure it, we must first see it, and before we see it, we must agree on what we are looking at.