Frontotemporal Dementia

Abstract

Frontotemporal dementia is a family of neurodegenerative syndromes that target the frontal and anterior temporal lobes, so the earliest losses are in social conduct, judgment, emotion, and language rather than in memory. It belongs to cognitive psychology because it is a slow lesion study of the circuits that support personality, self-regulation, and social cognition, exposing how the brain builds the faculties that make a person themselves. It is not one disease but a syndrome with a behavioural variant and two language variants, each tied to a distinct atrophy pattern, and beneath it lies a loosely coupled layer of molecular pathologies (abnormal tau, TDP-43, or FUS protein) and a strong genetic contribution. Beginning before age 65, it is the leading cause of dementia in younger adults, and the mismatch between syndrome, pathology, and gene is a central puzzle of neurodegeneration.

Keywords: frontotemporal dementia, behavioural variant, frontotemporal lobar degeneration, social cognition

Frontotemporal dementia (FTD) is a clinical term for a group of neurodegenerative syndromes in which degeneration concentrated in the frontal and temporal lobes produces early, progressive change in behaviour, personality, or language, while memory and perception are comparatively spared at first (Bang et al., 2015). The name locates the disease anatomically: the frontal lobes govern judgment, restraint, planning, and social behaviour, and the anterior temporal lobes anchor the meaning of words and objects, so degeneration here strikes precisely the faculties that a memory-first dementia leaves untouched until late. What makes FTD a subject for cognitive psychology, and not for neurology alone, is that its selective dissolution of these systems exposes the architecture of personality, self-control, and social understanding more sharply than almost any other condition (Olney et al., 2017).

What Frontotemporal Dementia Is

Frontotemporal dementia is a family of syndromes, not a single disease. A syndrome is a recognisable pattern of signs and symptoms; the disease is the specific pathological process that produces it. FTD names the clinical patterns — an insidious, progressive decline led by change in behaviour, personality, or language, with relative sparing of memory early on — while a group of related pathologies, collectively called frontotemporal lobar degeneration, lie beneath them (Neary et al., 1998), (Bang et al., 2015). Holding the syndrome apart from the disease is the single most important idea for understanding the condition, because it explains why two people with the same clinical diagnosis can have entirely different molecular illnesses, different genetics, and different prognoses.

The concept has a long history. In 1892 Arnold Pick described patients with progressive language and behavioural change whose brains showed circumscribed atrophy of the frontal and temporal lobes, and the lobar atrophy and the distinctive inclusion bodies later found in some such cases came to bear his name as Pick's disease (Bang et al., 2015). For much of the twentieth century these presentations were grouped loosely and often mistaken for Alzheimer's disease or for a psychiatric disorder. The modern framework arrived in 1998, when an international group set out consensus clinical criteria for frontotemporal lobar degeneration, distinguishing a frontal, behaviour-led presentation from temporal, language-led ones and giving the field a common vocabulary (Neary et al., 1998).

To qualify as FTD, a case must show a progressive, neurodegenerative deterioration led by the frontal or temporal syndrome rather than by amnesia, and exclusions rule out cases better explained by another neurological disease, a systemic illness, or a primary psychiatric disorder (Rascovsky et al., 2011). That last exclusion is unusually hard to apply, because the earliest behavioural changes of FTD — apathy, disinhibition, loss of empathy — closely mimic depression, mania, or a personality change, and distinguishing the two is one of the field's recognised diagnostic challenges (Ducharme et al., 2020).

Types of Frontotemporal Dementia

In the international MeSH vocabulary, the descriptor Frontotemporal Dementia is filed beneath *Frontotemporal Lobar Degeneration*, the broader pathological family, reflecting that the clinical dementia is one expression of that degeneration. MeSH indexes one more specific form directly under the descriptor: *Pick Disease of the Brain*, the historical eponym for cases with the characteristic lobar atrophy and Pick bodies (Table 1). Clinically, however, the field organises FTD around three prototypical variants defined by which faculty fails first, and these clinical variants cut across the molecular and eponymic categories rather than lining up with them one-to-one (Bang et al., 2015).

The *behavioural variant* (bvFTD) is the most common and is led by a progressive change in personality and social conduct — disinhibition, apathy, loss of empathy, compulsive or stereotyped behaviour, and altered eating — with atrophy concentrated in the medial and orbital frontal lobes and the anterior temporal lobes (Rascovsky et al., 2011). The two *language variants* are forms of primary progressive aphasia: the *semantic variant*, marked by fluent but empty speech and a loss of word and object meaning with anterior temporal atrophy, and the *nonfluent/agrammatic variant*, marked by effortful, agrammatic speech with left frontoinsular atrophy (Bang et al., 2015), (Olney et al., 2017). A substantial minority of patients also develop motor features overlapping with amyotrophic lateral sclerosis or with atypical parkinsonism, placing FTD at the head of a broad clinical spectrum (Convery et al., 2019).

Table 1

The MeSH Subtype and the Clinical Variants of Frontotemporal Dementia

Note. MeSH indexes Pick Disease of the Brain directly under the Frontotemporal Dementia descriptor; the behavioural and language variants are the standard clinical distinctions of the field (Rascovsky et al., 2011), (Bang et al., 2015). MeSH is an indexing classification, not a claim that these categories are mutually exclusive: the eponymic, clinical, and molecular schemes describe the same disease from different angles and cross-classify real cases. The parent pathological family, frontotemporal lobar degeneration, is the substrate common to all of them (Neary et al., 1998).

How the Frontal Networks Break Down

The faculties FTD attacks are not seated in single spots but distributed across large-scale brain networks, and this is the key to why its variants take the forms they do. The behavioural variant reflects degeneration of a *salience network* anchored in the anterior cingulate and frontoinsular cortex, which assigns emotional and social significance to events and helps regulate behaviour; the semantic variant reflects loss of the anterior temporal *semantic* hub; and the nonfluent variant reflects degeneration of a left frontal *speech-production* network (Seeley et al., 2009). A landmark finding is that each neurodegenerative disease appears to target a specific intrinsic network, spreading through its connections rather than striking tissue at random — so the clinical syndrome is the functional signature of the network that is failing (Seeley et al., 2009). The contrast with Alzheimer's disease is instructive: where the behavioural variant of FTD dismantles the salience network, Alzheimer's disease preferentially attacks the default mode network that supports episodic memory, and this divergence of target networks is part of why the two dementias produce such different first symptoms (Seeley et al., 2009), (Warren et al., 2013).

When degeneration begins in the salience network of the medial and orbital frontal lobes, the systems that restrain impulses, weigh social consequences, and generate empathy are attacked, and the result is the behavioural variant: a person who was reserved becomes tactless and disinhibited, or a person who was engaged becomes apathetic and indifferent, while their memory and perception remain intact enough that the change is read by families as a change in *who they are* (Rascovsky et al., 2011). Because the orbitofrontal cortex supports the evaluation of reward and the inhibition of prepotent responses, its degeneration also produces the compulsive rituals and the dietary changes — a craving for sweets, overeating — that form part of the diagnostic picture (Piguet et al., 2011). When degeneration begins instead in the anterior temporal lobe, the meaning hub is dismantled and the semantic variant follows; when it begins in the left frontoinsular speech network, the nonfluent variant follows (Bang et al., 2015).

The behavioural variant is also where FTD most clearly informs cognitive psychology, because its deficits fall on executive function and social cognition — the very faculties that let a person plan, inhibit, and model other minds. Patients show early failures of the theory of mind needed to infer others' beliefs and feelings, and a dysexecutive profile of impaired planning, mental flexibility, response inhibition, and working memory, even while episodic memory and visuospatial skills are comparatively preserved (Piguet et al., 2011). This pattern is the near-mirror image of the amnestic profile that defines early Alzheimer's disease, in which episodic memory fails first (McKhann et al., 2011), and the disorder thereby dissociates social and executive cognition from the memory systems that dominate other dementias.

The Cognitive Cost of Frontal Degeneration

The defining cost of FTD is not forgetting but the erosion of the faculties that regulate conduct and hold a personality together, and the *shape* of that cost differs sharply across variants. In the behavioural variant, the loss is of self-regulation and social cognition: the capacity to inhibit an impulse, to feel what another person feels, and to bring behaviour into line with social context. Because these depend on frontal networks that mature late and support the highest-order control of behaviour, their degeneration produces changes — disinhibition, apathy, loss of empathy — that families experience as a transformation of character rather than as a cognitive test failure (Rascovsky et al., 2011). Crucially, the person often lacks insight into the change, which both distinguishes FTD from a psychiatric disorder and compounds its burden (Ducharme et al., 2020).

In the semantic variant, the cost falls on semantic memory: because the anterior temporal lobe binds together the many features of a concept, its degeneration produces a loss of knowledge that extends beyond words to recognising objects, faces, and their uses (Bang et al., 2015). In the nonfluent variant, the cost is to the motor planning of speech and to grammar, so that the effort of producing and ordering words dominates while comprehension of single words is spared (Olney et al., 2017). What all three share is the early sparing of episodic memory and perception, the domains that define Alzheimer's disease, which is what allows FTD to be recognised as a distinct family rather than folded into a general dementia (Bang et al., 2015). As the underlying degeneration spreads, however, this selectivity fades and the syndromes converge into a broader decline.

The scale of the cost is amplified by *when* it strikes. FTD is the most common cause of dementia before the age of 65, with onset typically in the mid-50s to early 60s, so it removes judgment, restraint, and language from people at the height of their working and family lives, and its prevalence in the presenile range rivals that of early-onset Alzheimer's disease (Onyike & Diehl-Schmid, 2013).

Worked Example

The diagnosis of the behavioural variant is instructive because it turns a change in personality — the least quantifiable of clinical impressions — into a countable decision rule. The international consensus criteria define six core behavioural and cognitive features, and require that a threshold number be present for a diagnosis of *possible* bvFTD (Rascovsky et al., 2011). The six features are: (1) behavioural disinhibition; (2) apathy or inertia; (3) loss of sympathy or empathy; (4) perseverative, stereotyped, or compulsive behaviour; (5) hyperorality or dietary change; and (6) a dysexecutive neuropsychological profile with relative sparing of memory and visuospatial function. Let each feature be coded as present ($1$) or absent ($0$), and let $n$ be their sum. The rule for possible bvFTD is a simple threshold:

$$n = \sum_{i=1}^{6} f_i, \qquad \text{possible bvFTD} \iff n \ge 3$$

Consider a 58-year-old whose family reports that he has become socially tactless and makes inappropriate remarks (disinhibition, $f_1 = 1$), has lost interest in his hobbies and family (apathy, $f_2 = 1$), no longer seems moved by his wife's distress (loss of empathy, $f_3 = 1$), and has developed a craving for sweet foods (hyperorality, $f_5 = 1$), with normal memory on testing but poor planning and mental flexibility (dysexecutive profile, $f_6 = 1$). He shows no compulsive rituals ($f_4 = 0$). Summing:

$$n = 1 + 1 + 1 + 0 + 1 + 1 = 5 \ge 3 \;\Rightarrow\; \text{possible bvFTD}$$

The threshold structure repays careful reading because it marks its own limits. Requiring three of six features is a deliberate trade of sensitivity against specificity: too low a threshold would sweep in ordinary psychiatric illness and normal ageing, while too high a threshold would miss early cases (Rascovsky et al., 2011). The move from *possible* to *probable* bvFTD requires more than the behavioural count — it demands evidence of functional decline and imaging that shows frontotemporal atrophy or hypometabolism, precisely because the behavioural features alone overlap so heavily with primary psychiatric disorders (Ducharme et al., 2020). The rule is clean; the judgment it supports is probabilistic, and the criteria explicitly build in the imaging and functional evidence needed to raise confidence.

Discussion

The scientific value of FTD for cognitive psychology is that it dissociates faculties that a healthy mind fuses seamlessly. Where a memory-first dementia leaves social conduct and language intact until late, FTD does the reverse, and in doing so it shows that judgment, empathy, self-restraint, and word meaning are not diffuse properties of a general intelligence but products of specific, separable brain networks that can fail one at a time (Seeley et al., 2009). The behavioural variant in particular has become a central source of human evidence about the frontal control of behaviour and about social cognition, because it lesions those systems progressively and selectively in a way no experiment could ethically arrange (Piguet et al., 2011).

The three demonstrations on this page track that idea at three levels. The clinical-variant display shows where each variant's degeneration begins and which faculty that network supports; the diagnostic-threshold tool shows how a change in personality is converted into a countable criterion; and the molecular-pathology map shows how the clinical syndrome relates to the protein and gene beneath it. Where the account remains genuinely open is at the *boundaries between the levels*. The clinical variant, the molecular pathology, and the causative gene form three cross-cutting classifications that predict one another only imperfectly: the same behavioural syndrome can arise from tau or from TDP-43 pathology, and the same gene can produce different clinical pictures, so relating a living patient's syndrome to the disease beneath it is one of the field's central problems (Meeter et al., 2017), (Convery et al., 2019).

Cognitive Implications

The most far-reaching implication of FTD is that it decouples a *cognitive syndrome* from a *disease entity* at two levels at once — clinical from molecular, and molecular from genetic. Frontotemporal lobar degeneration comes in three broad molecular forms, defined by which protein aggregates abnormally in the brain: FTLD-tau, FTLD-TDP (built on the TDP-43 protein), and the rarer FTLD-FUS (Neumann et al., 2006). The discovery that ubiquitinated TDP-43 is the pathological protein in the largest group of cases unified a set of previously mysterious syndromes and linked FTD to amyotrophic lateral sclerosis, which shares the same pathology (Neumann et al., 2006). Yet the clinical variant does not read off the molecular subtype: a behavioural presentation may rest on tau or on TDP-43, and only the semantic variant maps reliably to one substrate.

The genetic layer sharpens the lesson. About a third of FTD is strongly heritable, and three genes account for most familial disease: *MAPT*, which encodes the tau protein and whose mutations cause a tau pathology (Hutton et al., 1998); *GRN* (progranulin), whose mutations cause a TDP-43 pathology (Baker et al., 2006); and *C9orf72*, whose hexanucleotide repeat expansion is the most common genetic cause of both FTD and ALS and also produces TDP-43 pathology (DeJesus-Hernandez et al., 2011), (Renton et al., 2011). The mapping from gene to protein is tight — *MAPT* to tau, *GRN* and *C9orf72* to TDP-43 — but the mapping from either to the clinical syndrome is loose, so a family carrying one mutation can show behavioural, language, and motor presentations across its members. For the science of cognition, this triple dissociation is a reminder that the route from molecule to mind runs through the *anatomy of function*: it is *where* a disease starts in the brain's networks, more than *what* the disease is molecularly, that determines the shape of the mind's decline.

Current Directions

Three fronts are active. The first is *biomarkers that identify the molecular pathology during life*. Because the clinical syndrome predicts the underlying protein only imperfectly, the field is pursuing fluid and imaging markers — blood and cerebrospinal-fluid measures of neurodegeneration, and markers that distinguish tau from TDP-43 pathology — so that a living patient can be assigned to a molecular subtype without waiting for autopsy (Meeter et al., 2017). This matters increasingly because gene-specific therapies are entering trials, and they require knowing which pathology and which gene are at work (Boeve et al., 2022).

The second front is *genetic FTD as a window for prevention*. Because *MAPT*, *GRN*, and *C9orf72* mutations are highly penetrant, carriers can be identified before symptoms begin, and large longitudinal studies now track them through the presymptomatic phase to define the earliest biological and cognitive changes and to time intervention (Boeve et al., 2022). Progranulin deficiency in particular is a rational drug target, since *GRN* mutations act by reducing the protein, and strategies to restore it are in development (Baker et al., 2006). The third front is *diagnosis at the psychiatric boundary*. Because early bvFTD so closely mimics depression, mania, and personality change, work continues on the features, imaging, and biomarkers that separate a neurodegenerative disorder from a primary psychiatric one, so that patients are neither misdiagnosed for years nor wrongly labelled (Ducharme et al., 2020).

Common Misconceptions

Frontotemporal dementia is a kind of Alzheimer's disease.

No. FTD is a distinct family of syndromes with a different anatomy (frontal and temporal rather than hippocampal), a different age of onset (typically before 65), and different pathologies (tau, TDP-43, or FUS rather than amyloid and tau). Early memory loss points away from FTD, not toward it (Bang et al., 2015).

Dementia always begins with memory loss.

In FTD, memory is comparatively spared early, and the first changes are in behaviour, personality, or language. This is exactly why FTD is so often missed or misdiagnosed at first — it does not fit the memory-first template that the word *dementia* evokes (Rascovsky et al., 2011).

The behavioural changes are just depression or a midlife crisis.

The apathy, disinhibition, and loss of empathy of early bvFTD closely mimic psychiatric illness, and telling them apart is a recognised diagnostic challenge. But in FTD the changes are driven by progressive brain degeneration, usually without insight, and are accompanied over time by atrophy on imaging (Ducharme et al., 2020).

Frontotemporal dementia is purely a genetic disease.

Only about a third of cases are strongly heritable. The majority are sporadic, with no clear family history. Where disease is inherited, mutations in MAPT, GRN, and C9orf72 account for most of it, but most patients carry none of these (Convery et al., 2019).

Glossary

Anterior temporal lobe.

The front portion of the temporal lobe, including the temporal pole; degenerates in the semantic variant and acts as a hub binding the features of word and object meaning.

Behavioural variant (bvFTD).

The most common form of FTD, led by progressive change in personality and social conduct — disinhibition, apathy, loss of empathy, compulsive behaviour, and dietary change — with frontal and anterior temporal atrophy.

C9orf72.

A gene whose hexanucleotide repeat expansion is the most common genetic cause of frontotemporal dementia and of amyotrophic lateral sclerosis; produces TDP-43 pathology.

Disinhibition.

A failure to restrain impulses or bring behaviour into line with social context; a core early feature of the behavioural variant, reflecting orbitofrontal degeneration.

Executive function.

The set of control processes — planning, mental flexibility, and response inhibition — supported by the frontal lobes; impaired early in the behavioural variant.

Frontotemporal lobar degeneration.

The family of neurodegenerative pathologies, defined by abnormal tau, TDP-43, or FUS protein, that underlies the clinical frontotemporal dementia syndromes; the parent category in MeSH.

GRN (progranulin).

A gene whose loss-of-function mutations reduce the progranulin protein and cause a TDP-43 frontotemporal pathology; a rational target for protein-restoring therapy.

Hyperorality.

A change in eating behaviour — overeating, a craving for sweets, or putting non-food items in the mouth — that forms one of the six core diagnostic features of the behavioural variant.

MAPT.

The gene encoding the tau protein; its mutations cause an inherited tau frontotemporal pathology and were the first FTD gene identified.

Nonfluent/agrammatic variant.

The language variant of FTD marked by effortful, halting, agrammatic speech with left frontoinsular atrophy; a form of primary progressive aphasia.

Pick's disease.

The historical eponym, and the MeSH subtype, for frontotemporal cases with circumscribed lobar atrophy and tau-positive Pick bodies; named for Arnold Pick.

Salience network.

A large-scale brain network anchored in the anterior cingulate and frontoinsular cortex that assigns emotional and social significance to events; its degeneration underlies the behavioural variant.

Semantic memory.

The store of general knowledge about words, objects, and concepts; progressively lost in the semantic variant as the anterior temporal hub degenerates.

TDP-43.

A protein that aggregates abnormally in the largest molecular subgroup of frontotemporal lobar degeneration and in most amyotrophic lateral sclerosis, linking the two diseases.

Theory of mind.

The capacity to infer the beliefs, intentions, and feelings of others; impaired early in the behavioural variant as social-cognitive networks degenerate.

Key Researchers

John R. Hodges

Neurologist at the University of Sydney, formerly Cambridge, who helped define the frontotemporal syndromes and co-authored both the revised behavioural-variant diagnostic criteria and the clinical-staging framework that organise the field. Wikidata

Bruce L. Miller

Behavioral neurologist at the University of California, San Francisco who has led the modern clinical characterisation of frontotemporal dementia, including the large-scale network-degeneration account and the major clinical reviews. ORCID

Arnold Pick

Czech-German psychiatrist and neurologist at the German University of Prague who, in the 1890s, described the circumscribed lobar atrophy and progressive behavioural and language change now recognised as part of the frontotemporal spectrum; his name survives in Pick's disease and Pick bodies. Wikipedia

Jonathan D. Rohrer

Clinical neurologist at University College London who leads longitudinal genetic-FTD cohort work and investigates the genetics, imaging, and fluid biomarkers that are reshaping diagnosis and trials. ORCID

Julie S. Snowden

Neuropsychologist at the University of Manchester who helped define the frontotemporal lobar degeneration syndromes and their consensus clinical criteria and co-authored the discovery of progranulin as a cause of the disease. ORCID

Frequently Asked Questions

What is frontotemporal dementia?

It is a family of neurodegenerative syndromes in which degeneration of the frontal and anterior temporal lobes produces early, progressive change in behaviour, personality, or language, while memory and perception are comparatively spared at first. It is the leading cause of dementia before the age of 65.

How is it different from Alzheimer's disease?

Alzheimer's disease usually begins with memory loss and targets the hippocampus and posterior brain, typically after 65. Frontotemporal dementia begins with changes in conduct or language, targets the frontal and temporal lobes, and typically starts earlier, in the 50s and 60s. Their underlying pathologies also differ.

What are the main variants?

A behavioural variant, led by change in personality and social conduct such as disinhibition, apathy, and loss of empathy; and two language variants. The semantic variant is marked by loss of word and object meaning, and the nonfluent or agrammatic variant by effortful, halting speech.

What causes it?

Beneath the clinical syndromes lies frontotemporal lobar degeneration, which comes in molecular forms defined by abnormal tau, TDP-43, or FUS protein. About a third of cases are inherited, most often through mutations in the genes MAPT, GRN, and C9orf72; the rest are sporadic.

Is it inherited?

Sometimes. Roughly a third of people with FTD have a strong family history, and highly penetrant mutations in MAPT, GRN, or C9orf72 explain most inherited cases. The majority of cases, however, occur without a clear family history.

Why is it so often misdiagnosed?

Because its earliest signs, such as apathy, disinhibition, and loss of empathy, closely resemble depression, mania, or a personality change, and because it does not begin with the memory loss most people associate with dementia. Distinguishing early behavioural-variant FTD from a primary psychiatric disorder is a recognised diagnostic challenge.

Can it be treated?

No treatment yet halts the underlying degeneration. Management centres on treating symptoms, supporting carers, and safety, and gene-specific therapies for inherited forms are now entering clinical trials. Accurate diagnosis of the syndrome, pathology, and any causative gene increasingly guides care.

Does it affect memory?

Episodic memory is comparatively spared in the early years, which is one of the features that distinguishes FTD from Alzheimer's disease. The semantic variant does erode semantic memory, the knowledge of the meaning of words and objects, and as the disease advances, a broader decline including memory usually follows.

Support Organizations

References