Elucidating the Neurobiological Underpinnings of Mild Behavioral Impairment in Tauopathies: Clinical and Molecular Insights
11st Department of Neurology, Aiginition University Hospital, Medical School, National and Kapodistrian University of Athens, 11528 Athens, Greece; angelthal@med.uoa.gr (E.A.); jpapatriantafyllou@gmail.com (J.P.); sokpapa@med.uoa.gr (S.P.)
2Third Age Day Care Center IASIS, 16562 Athens, Greece
3School of Medicine and Surgery, University of Milano-Bicocca, 20900 Monza, Italy
*Correspondence: chiara.villa@unimib.it; Tel.: +39-02-6448-8138Abstract
Mild behavioral impairment (MBI) is a clinical syndrome characterized by the late-life onset and persistence of neuropsychiatric symptoms (NPSs), representing a change from longstanding behavior or personality and considered a potential prodrome of neurodegenerative disease. MBI is classified into five domains: decreased motivation, affective dysregulation, impulse dyscontrol, social inappropriateness, and psychotic symptoms. In this narrative review, we synthesize clinical, neuroanatomical, and molecular evidence linking MBI to the spectrum of tauopathies, including Alzheimer’s disease (AD), frontotemporal spectrum disorders (FTSDs), and primary four-repeat tauopathies such as progressive supranuclear palsy (PSP) and corticobasal degeneration (CBD). Emerging evidence suggests that early behavioral symptoms associated with MBI may reflect the selective vulnerability of frontolimbic, salience, default mode, and frontostriatal networks to tau-mediated neurodegeneration. Mechanistically, converging findings support roles for tau-related synaptic dysfunction, including synaptotoxic soluble tau species, cytoskeletal and axonal transport disruption, monoaminergic neurotransmitter imbalance in brainstem systems, and neuroinflammatory and glial pathways. We also highlight genotype-related behavioral profiles in genetic frontotemporal lobar degeneration and discuss how scalable blood-based biomarkers, including neurofilament light chain, glial fibrillary acidic protein, and plasma phospho-tau species, may complement MBI-based phenotyping for differential diagnosis and prognostic stratification in clinical research.
1. Introduction
Mild behavioral impairment (MBI) is a syndrome characterized by later-life onset and the persistence of neuropsychiatric symptoms (NPSs), representing a potential early manifestation of neurodegenerative diseases [1,2]. The concept of MBI was developed to capture the neurobehavioral axis of prodromal neurodegeneration that can emerge with or without objective cognitive impairment, complementing the construct of mild cognitive impairment (MCI) [1,2]. NPSs, such as depressive symptoms, anxiety, apathy, irritability, loss of empathy and psychotic manifestations, are common across neurodegenerative disorders and are increasingly recognized as an integral part of the disease, potentially representing early clinical signs of the underlying neurodegenerative process [3]. In this regard, MBI provides a standardized framework to appropriately capture these early behavioral changes [2]. The recent MBI criteria highlight some core features that help distinguish MBI-related NPSs from NPSs in dementia or primary psychiatric conditions, emphasizing the onset in later life, the persistence, and the clear change from longstanding baseline behavior, thereby improving the signal-to-noise ratio when studying prodromal disease [2].
Tauopathies comprise a heterogeneous group of neurodegenerative diseases defined by the abnormal aggregation of the tau protein, spanning Alzheimer’s disease (AD) and primary tauopathies such as progressive supranuclear palsy (PSP), corticobasal degeneration (CBD), and frontotemporal lobar degeneration with tau pathology (FTLD-tau) [4]. Despite their marked clinicopathological overlap, tauopathies show, at least partially, some distinct patterns of network degeneration and neuropsychiatric phenotypes, including apathy, affective dysregulation, impulsivity and disinhibition, as well as changes in social cognition [4], which may potentially map onto MBI domains. The concept of domain-based mapping is attractive because it allows behavioral symptoms to be studied as structured clinical outcomes linked to large-scale neural networks, rather than as nonspecific psychiatric comorbidities. Importantly, the current biomarker era is altering how prodromal neurodegenerative syndromes are defined and studied. While plasma, cerebrospinal fluid (CSF) and neuroimaging tools are rapidly evolving for AD, biomarker translation for primary tauopathies remains a critical unmet need, particularly for participant enrichment in clinical trials at the early stages of the diseases [4]. Recent translational advances, including tau species that may support discrimination among primary tauopathies in CSF, highlight the importance of a biologically based diagnosis. Within this emerging landscape, MBI assessment could function as a scalable clinical tool that could complement biomarker strategies by identifying individuals who might be more likely to harbor evolving neurodegenerative pathology. Emerging evidence also suggests that specific MBI domains may show longitudinal associations with plasma markers relevant to AD pathophysiology, including phosphorylated tau and potentially broader metabolic or neurodegenerative signatures [5]. Although the potential underlying pathophysiology of MBI in AD has been already reviewed [5,6,7], whether and how these relationships can be generalized, especially across the whole spectrum of tauopathies, remains largely underexplored.
In this narrative review, given the growing evidence of MBI in prodromal neurodegeneration, we examine MBI as a transdiagnostic construct across tauopathies, focusing on the clinician and mechanistic links between tau-mediated network degeneration and emergent behavioral changes. For this purpose, we synthesize current knowledge on MBI across the spectrum of tauopathies, encompassing clinical, neuroanatomical, and molecular insights, highlighting both shared and disease-specific features, aiming to position MBI as a useful link between clinical phenotyping and molecular neurodegeneration.
2. Methods
To identify relevant evidence, we followed a structured literature search approach. We searched MEDLINE and Scopus databases with no a priori time restrictions, using combinations of controlled vocabulary terms and free-text keywords related to the following keywords: “mild behavioral impairment”, “neuropsychiatric symptoms”, individual MBI domains including “apathy”, “depression”, “anxiety”, “irritability”, “agitation”, “disinhibition”, “compulsivity”, “psychosis”, “hallucinations”, “delusions”, and tau-related terms, including “tauopathy”, “tau proteinopathy”, “tau”, “Alzheimer’s disease”, “frontotemporal dementia”, “behavioral variant FTD”, “progressive supranuclear palsy”, “corticobasal degeneration”, and “Pick’s disease”. Search results were screened at the title and abstract level, and potentially relevant full texts were reviewed. We primarily included clinical studies investigating MBI in tauopathies using MBI-C or compatible operational definitions, as well as preclinical studies elucidating potential mechanisms linking tau-mediated network or cellular dysfunction to behavioral phenotypes consistent with MBI domains, including synaptic dysfunction, neurotransmitter alterations and neuroinflammatory pathways. For discussion purposes, we included additional studies reporting NPSs in the broader context of preclinical or prodromal stages of neurodegenerative diseases when directly informative for tauopathy–MBI relationships. The reference lists of the included articles were manually screened with a snowball process to identify additional eligible studies. Given the expected heterogeneity in designs, instruments, and outcomes, evidence was synthesized narratively, organized in two levels: (i) tauopathy clinical syndrome; (ii) convergent molecular and network-related mechanisms linking tau pathology to early behavioral changes.
3. MBI Definition
MBI is a neurobehavioral syndrome proposed to capture the late-life emergence of persistent NPSs as an at-risk state for incident cognitive decline and dementia, with or without concurrent cognitive complaints. Although the concept has earlier roots in observations of prodromal behavioral change particularly in frontotemporal spectrum disorders (FTSDs), the recent research framework for MBI was defined by the International Society to Advance Alzheimer’s Research and Treatment: Alzheimer’s Association (ISTAART-AA), Neuropsychiatric Syndromes Professional Interest Area (PIA) in 2016 [2]. According to these criteria, MBI is characterized by the onset of behavioral or personality changes present, for at least 6 months, after the age of 50 years; the behavioral changes must represent a deviation from an individual’s usual behavior or personality and must not be better explained by a primary psychiatric diagnosis, a medical condition or substance effects [2].
Importantly, MBI may coexist with normal cognition, subjective cognitive decline (SCD), or MCI, but it is conceptually distinct from MCI and should not be viewed as a competing construct [2]. NPSs in the context of MBI must be directly associated with at least mild functional impairment, though not severe enough to meet the diagnostic criteria for dementia. MBI should be distinguished from both classical psychiatric disorders (major depression, bipolar disorder, general anxiety disorder, delusional disorder, schizophrenia, etc.) and the traditional behavioral and psychological symptoms of dementia (BPSD), as seen in established dementia [2]. MBI can be considered as the predementia behavioral axis that parallels the cognitive axis of neurodegeneration that is mainly captured by MCI [1,2].
A key strength of the MBI construct lies in its attempt to improve the specificity of the assessment of NPSs in older adults. Conventional neuropsychiatric frameworks often capture symptoms that are transient, reactive, or attributable to life stressors, whereas MBI requires later-life onset, persistence, and clear change from baseline. This distinction is particularly important in aging populations, in whom depression, anxiety, irritability, apathy, or suspiciousness may arise for multiple reasons unrelated to neurodegeneration. By requiring persistence for at least 6 months and excluding longstanding or recurrent psychiatric syndromes, MBI enriches for behavioral symptoms more likely to reflect evolving brain disease rather than nonspecific psychological distress [2].
The MBI syndrome is organized into five domains: (a) decreased motivation (apathy, reduced initiative), (b) emotional dysregulation (anxiety, depression, mood lability), (c) impulse dyscontrol (agitation, aggression, compulsivity), (d) social inappropriateness (disinhibition, lack of empathy), and (e) abnormal perception or thought content (hallucinations, delusions) [2]. This domain-based approach is especially relevant for tauopathies where different patterns of frontolimbic, salience network, and frontostriatal involvement might potentially underly distinct MBI presentations [8,9]. The operationalization of MBI in research and clinical settings has been greatly facilitated by the development of the MBI Checklist (MBI-C) [1]. The MBI-C is a 34-item instrument specifically developed for predementia populations, designed to directly align with the ISTAART-AA criteria. This is a major methodological advantage over instruments such as the Neuropsychiatric Inventory (NPI), originally developed for dementia populations with shorter reference windows [1]. MBI-C can potentially be used to detect behavioral changes across a wide range of neurodegenerative disorders. For instance, in Parkinson’s disease (PD), MBI-C was recently revised into a 24-item model to prevent floor effects and better capture the smallest manifestations of MBI in PD [10]. Behavioral abnormalities and cognitive impairments, especially executive dysfunctions, can appear early in amyotrophic lateral sclerosis (ALS) and are linked with faster disease progression, lower survival, and increased caregiver burden [11]. Also in this context, MBI was revealed to be a reliable clinical marker of cognitive deterioration both in ALS and other motor neuron diseases (MNDs), showing great potential for identifying dementia in its prodromal or preclinical stages [12,13].
It has been demonstrated that the MBI-C can be utilized across different clinical and research settings. In cognitive clinic cohorts, higher MBI-C scores have been associated with lower cognitive performance, and MBI prevalence increases across SCD, MCI, and dementia groups [14]. In SCD populations, the MBI-C was able to detect subtle but clinically relevant NPSs, and its total score showed significant associations with established neuropsychiatric and depressive symptom measures, including the NPI-Questionnaire (NPI-Q) and the Geriatric Depression Scale-15 (GDS-15), supporting its construct validity in early, predementia stages [15]. Collectively, these findings support the use of the MBI-C as the current standard instrument for MBI ascertainment in both research and increasingly in prodromal clinical phenotyping.
Growing longitudinal evidence further supports the prognostic significance of MBI. In older adults without dementia, MBI has been associated with worse baseline cognition, faster cognitive decline, and increased risk of incident dementia [16]. In a prospective Asian cohort, 38.6% of individuals with MBI developed dementia compared with 12.3% of those without MBI, corresponding to a 2.56-fold increased risk of incident dementia [16]. Similarly, in older adults without dementia from ADNI, MBI was associated with poorer cognition, greater and faster increase in amyloid burden, and an elevated risk of cognitive decline (HR 2.42) [17]. These data suggest that MBI can be a clinically meaningful prognostic construct linked to adverse cognitive trajectories.
Some evidence has also shown connections between genetic background and MBI, indicating that certain MBI domains correlate with different biochemical pathways involved in AD and other dementias. The possibility that MBI domains represent clinically apparent manifestations of underlying genetic susceptibility is highlighted by associations with some genetic variants (Table 1). Among them, the presence of the Apolipoprotein E (APOE) ε4 allele, the strongest common genetic risk factor for late-onset AD, was associated with a higher likelihood of affective dysregulation [18]. The same study also reported that variants located on MS4A4A and MS4A6A genes have been inversely associated with affective dysregulation, while ZCWPW1 variants have been related to decreased social inappropriateness and psychotic manifestations [18]. BIN1 and EPHA1 have been linked to psychosis, whereas NME8 has been inversely associated with apathy [18]. Other evidence links the presence of least one Met allele in the brain-derived neurotrophic factor (BDNF) gene with a higher likelihood of MBI in PD patients, as assessed by MBI-C [19].
Finally, recent work suggests that the way in which MBI is operationalized matters substantially for prognostic accuracy. Comparisons of different case definitions indicate that persistence-based definitions of late-life emergent NPSs outperform transient or nonspecific symptom definitions in dementia risk modeling, reinforcing one of the central conceptual premises of MBI, according to which, sustained behavioral change in later life is more informative than isolated NPSs when attempting to identify prodromal neurodegenerative diseases [20]. The MBI-C has already been translated and validated in multiple languages and settings, supporting its cross-cultural applicability and broader use in both research and clinical practice [21]. In addition, it has shown feasibility for remote administration, including telephone-based assessment in SCD cohorts, further supporting its practicality as a scalable screening tool beyond specialist memory clinics [15]. For these reasons, MBI is increasingly viewed as a scalable clinical enrichment construct that may complement biomarkers in identifying individuals at heightened risk for neurodegenerative disorders, including tauopathies.
4. MBI in the Spectrum of Tauopathies: Clinical Insights
4.1. MBI in Alzheimer’s Disease: Focusing on Tau Pathology
AD is the most extensively studied tauopathy in relation to MBI and currently provides the strongest evidence linking MBI symptoms with molecular markers of neurodegeneration. AD is characterized by the abnormal accumulation of tau as neurofibrillary tangles (NFTs) and neuropil threads mainly in neurons, as well as in dystrophic neurites accompanied by the aberrant deposition of amyloid-beta plaques extracellularly [22]. While AD has traditionally been conceptualized through its amnestic cognitive presentation, NPSs may emerge early in the disease course, sometimes even before prominent cognitive decline. In biomarker-defined amnestic MCI due to AD (aMCI-AD), approximately half of participants met criteria for MBI (48.4%; 30/62). Compared to 50 cognitively normal older adults, the aMCI-AD group was associated with higher total MBI severity, with the strongest effects in affective dysregulation, apathy, and impulse dyscontrol, whereas social inappropriateness and psychotic symptoms were not significantly different from cognitively normal controls [23].
In larger non-dementia AD spectrum cohorts, MBI has also been associated with poorer cognition and greater pathological burden. In ADNI participants without dementia, MBI was associated with worse global cognition and higher β-amyloid burden. Importantly, amyloid at least partially mediated the association between MBI and cognition, accounting for about 17% of the effect on global cognition and 38% for memory, executive, and language domains [17]. Longitudinally, MBI has been associated with more rapid amyloid accumulation and an elevated risk of cognitive decline [17]. Notably, new tau-focused data suggest that the association is not limited to amyloid. In Aβ-positive ADNI participants with normal cognition or MCI, MBI was associated with greater tau-positron emission tomography (PET) uptake in Braak I and Braak III areas, supporting a link between MBI and early cortical tau deposition in AD-vulnerable regions [24].
Blood-based biomarker studies point in the same direction. In ADNI participants with normal cognition or MCI, MBI was associated cross-sectionally with approximately 8% higher plasma p-tau181 and longitudinally with persistently higher p-tau181, along with decline in memory and executive function [25]. Survival analyses showed a 3.92-fold greater dementia incidence in those with MBI, whereas transient NPSs not meeting MBI criteria were not significantly different from the no-NPS group [25]. Similarly, in the NOLAN cohort, the MBI domain of psychotic features was associated with steeper increases in plasma pTau181 over one year, further supporting the role of tau-mediated pathways underlying MBI within the AD continuum [26]. Findings from CSF studies reinforce this distinction. Compared with no NPSs or transient NPS, MBI has been associated with lower CSF Aβ42 and Aβ42/40 ratios, higher p-tau and t-tau levels, and higher p-tau/Aβ42 and t-tau/Aβ42 ratios, while also conferring substantially greater risk of progression to dementia [27]. Accordingly, MBI was also investigated within the context of the A/T/N framework of AD. In a larger dementia-free cohort, individuals with MBI were correlated with CSF amyloid and tau positivity, as well as with an AD continuum biomarker profile, but not with non-AD pathology [28]. These observations further support the notion that persistent late-life emergent behavioral symptoms are more tightly linked to core AD pathophysiology than nonspecific or fluctuating NPS.
Interestingly, MBI can be proposed as the clinical manifestation of network-level consequences of tau deposition in AD. In this context, recent evidence in participants with aMCI and mild AD suggests that tau positivity might be linked to MBI indirectly through reduced salience network segregation, rather than through a simple direct biomarker-to-symptom association [9]. These findings suggest that emergent behavioral symptoms may arise when tau-related pathology disrupts large-scale functional systems involved in emotional salience, behavioral switching, and socioemotional regulation. This framework is consistent with the broader resting-state neuroimaging literature, demonstrating that AD preferentially may affect the default mode, salience, and limbic networks early in the disease course [29]. These networks are critically involved not only in memory and self-referential processing, but also in affective appraisal, motivation, and behavioral control. Collectively, these findings support a systems model in which MBI might represent the behavioral expression of selective network vulnerability within the AD continuum, linking regional tau pathology with NPS changes.
Importantly, emerging data suggest that different MBI domains may reflect distinct but overlapping pathogenic processes. In this regard, the MBI domain of psychotic manifestations may track tau-related change more closely, whereas decreased motivation and impulse dyscontrol may index broader metabolic and systems-level dysregulation [26]. Specifically, abnormal perception was associated with steeper increases in plasma pTau181 as mentioned above, whereas decreased motivation, and impulse dyscontrol were associated with homocysteine or insulin dysregulation [26]. Likewise, apathy might be particularly relevant in relation to tau-linked AD biomarker burden [30]. In particular, in older adults without dementia, the MBI domain of apathy was associated cross-sectionally and longitudinally with higher CSF p-tau181/Aβ42 and t-tau/Aβ42 ratios, and exploratory analyses also linked apathy with higher p-tau181 over a years-long follow-up [30]. Collectively, these findings support the view that MBI might be a biologically heterogeneous construct in which individual domains may reflect partially distinct—yet intersecting—molecular- and network-level mechanisms, reinforcing the value of domain-informed behavioral phenotyping for linking emergent NPSs to underlying AD pathophysiology.
Overall, MBI might represent a marker of higher amyloid burden, early tau accumulation, plasma p-tau181 elevation, and greater risk of clinical progression in AD [5,6]. This makes MBI particularly relevant in the AD continuum, where it may help identify patients with biologically active disease at a stage when behavioral change is already detectable but overt dementia has not yet developed.
4.2. MBI in Frontotemporal Spectrum Diseases Associated with FTLD-Tau Pathology
The behavioral variant FTD (bvFTD) is one of the most common clinical presentations of frontotemporal lobar degeneration (FTLD), representing the clearest illustration of the MBI construct in a non-AD context. Unlike AD, where memory loss typically dominates early presentations, bvFTD often begins with profound changes in behavior, personality, and social conduct that progressively worsen over time. These behaviors often emerge before measurable cognitive impairment, providing a strong rationale for considering MBI-like features as early indicators of bvFTD. This framework is particularly relevant in FTLD-tau, including Pick’s disease, the prototypical sporadic 3R tauopathy classically associated with early behavioral and socioemotional change. Notably, the five MBI domains map closely onto the core behavioral symptom clusters embedded in the FTDC criteria for possible bvFTD (disinhibition, apathy/inertia, loss of sympathy/empathy, perseverative/compulsive behaviors, hyperorality and dietary changes), supporting MBI as a structured syndromic lens for the earliest bvFTD phenotype [31].
Importantly, Taragano’s earlier work on late-life emergent behavioral changes without dementia helped motivate the later MBI framework [32,33,34]. According to Taragano and colleagues, MBI could be proposed to consist of four main features: (a) persistent behavioral alterations and mild psychiatric symptoms, particularly disinhibition; (b) absence of serious cognitive complaints; (c) normal daily functioning; (d) absence of dementia. In an early prospective cohort study by Taragano and colleagues, individuals with MBI showed a high risk of progression to dementia, including FTD and AD [32]. In another longitudinal cohort of individuals presenting with MBI, conversion occurred predominantly to bvFTD, with fewer AD converters [35]. Baseline executive dysfunction, severe impairment in theory of mind, and increased frontal atrophy were associated with higher conversion risk [35]. Furthermore, MBI has been associated with a higher risk of incident dementia, compared to the risk observed in a primary psychiatric comparison group, supporting the value of MBI in recognizing early, prodromal stages of FTD [36]. These findings support the view that incorporating social–cognitive testing and targeted frontal MRI readouts into MBI assessments may improve prognostic enrichment and reduce misclassification with primary psychiatric presentations.
Currently, the NPI, the NPI-Q and the Frontal Behavioral Inventory (FBI) are among the most widely used tools for evaluating neuropsychiatric symptoms in bvFTD [37,38]. However, the MBI and MBI-C are attracting increasing interest in assessing behavioral changes in patients with suspected bvFTD, especially at the early stages before the onset of prominent dementia [39]. Compared with the NPI or NPI-Q, which are typically anchored to shorter symptom windows and dementia populations, the MBI-C enforces persistence (≥6 months) and late-life onset, which may improve specificity for neurodegeneration-related behavioral change in bvFTD spectrum presentations [1]. One of the first cases of the application of the MBI-C in a patient with MBI who eventually developed bvFTD was published in 2018, where the MBI-C demonstrated better ability in identifying impulse dyscontrol symptoms compared to the NPI-Q [40].
The first study that investigated the application of the MBI-C in patients with bvFTD was by Cui and colleagues in 2023 [39]. In this Chinese cohort, including 52 patients with bvFTD and 82 healthy controls, MBI-C was greater than 0 in all patients with bvFTD and in about 40% of healthy controls [39]. The optimal cutoff point of the MBI-C for discriminating patients with bvFTD from healthy controls was 5.5, with 100% sensitivity and 83% specificity [39]. On the other hand, the NPI-Q for this discrimination had lower sensitivity and specificity measures compared to the MBI-C [39], suggesting that the MBI-C might be a more sensitive and specific tool for bvFTD compared to the NPI-Q. The optimal cutoff point of the Chinese version of the MBI-C in identifying AD dementia was 6.5 [41]. In accordance, a recent study indicated that the optimal cutoff point for distinguishing MCI due to AD from healthy controls in the Greek population was 9.5 [21]. A plausible explanation for the lower optimal MBI-C cutoff in bvFTD compared with AD is that behavioral and personality changes are the core, early features of bvFTD, so even a modest symptom burden might more reliably distinguish patients with bvFTD from controls without bvFTD. In contrast, NPSs in AD dementia might be more heterogeneous in profile and timing, often emerging alongside broader cognitive decline, which can shift the threshold needed to maximize discrimination. Moreover, a substantial proportion of healthy controls may score above zero on the MBI-C, so a somewhat higher cutoff in AD studies may be required to limit false positives. Furthermore, Dodich et al. showed that patients with bvFTD exhibited more prominent and more specific theory-of-mind impairment compared to AD [42], suggesting that early social inappropriateness in MBI might reflect disease-specific disruption of socioemotional inference rather than generalized dementia severity. Hence, in bvFTD, even modest behavioral changes might be diagnostically informative at an earlier point; meanwhile, in AD, a higher overall symptom burden may be needed to optimize discrimination.
In the study by Cui and colleagues, in patients with bvFTD, apathy was the most common MBI domain, followed by impulse dyscontrol, affective dysregulation, social inappropriateness and psychosis [39]. In subgroup analyses, impulse dyscontrol and apathy were the most prevalent MBI domains in mild and moderate-to-severe FTD, respectively [39]. The relative frequency of the MBI domains in bvFTD is broadly consistent with the clinical prominence of apathy, disinhibition, loss of empathy, and compulsive or perseverative behaviors in bvFTD as reflected in FTDC-based literature [31].
Regarding the neuroimaging correlates, apathy in bvFTD has been associated with atrophy in midline prefrontal regions, including areas of the orbitofrontal cortex, anterior/dorsal cingulate cortex [43], as well as fractional anisotropy changes in inferior fronto-occipital fasciculus and forceps minor [44]. Disinhibition has been linked to altered fiber integrity in the superior longitudinal fasciculus [44]. Moreover, Whitwell and colleagues demonstrated substantial neuroanatomical heterogeneity in bvFTD, identifying four distinct MRI-based atrophy subtypes (frontal-dominant, frontotemporal, temporal-dominant, and temporofrontoparietal). Notably, these subtypes differed on cognitive measures, including episodic memory, executive function, and confrontation naming, but showed comparable behavioral severity as being indexed by the NPI, suggesting that a similar overall NPS burden may accompany divergent regional patterns of neurodegeneration [45]. This anatomical heterogeneity supports domain-based behavioral phenotyping as a complementary approach to syndrome labels, because similar overall NPS burden may arise from divergent network-level degeneration patterns. However, it would be interesting for future studies to investigate the underlying neuroanatomical and functional correlations of MBI domains in bvFTD, especially at the earliest stages, where the sensitivity of MBI-C might be higher in detecting subtle cases.
In bvFTD, impairment in social cognition is a core early feature that overlaps substantially with the MBI domain of social inappropriateness. Rankin and colleagues showed that patients with FTLD demonstrate distinct disturbances in empathy [46,47], supporting the view that empathy loss reflects primary disease-related disruption of socioemotional processing. Consistent with this, the international consensus recommendations on differentiating bvFTD from primary psychiatric disorders emphasize that at least one formal social cognition test should be included in the standard assessment [47], as social–cognitive deficits might be considered particularly informative when behavioral symptoms dominate and cognitive screening may still be relatively preserved. Given that deficits in empathy and social cognition often precede overt cognitive impairment, the NIC-FTD consensus recommends incorporating at least one formal social cognition measure into the routine assessment when bvFTD is suspected, an approach that aligns directly with the MBI domain of social inappropriateness and may reduce psychiatric misdiagnosis. Together, these data reinforce the clinical value of systematically capturing MBI “social inappropriateness” features in the suspected early stages of bvFTD, both to support earlier detection and to reduce misdiagnosis as a primary psychiatric condition.
In the study by Cui and colleagues, all MBI domains, with the exception of psychotic features, were more common in patients with bvFTD compared to healthy controls. Interestingly, the prevalence of psychosis was similar between people with mild bvFTD and healthy controls, whereas it was higher in the subgroup of people with moderate-to-severe bvFTD [39]. This pattern suggests that most MBI domains are sensitive to early bvFTD, whereas psychotic symptoms might usually emerge later in the disease course, and therefore may be less useful for early-stage discrimination. Importantly, no significant associations were observed between the MBI-C and cognitive scales in the study by Cui and colleagues, including both global (MMSE, MoCA) and domain-specific tests (Auditory Verbal Learning Test (AVLT) learning, AVLT Delayed recall, AVLT Cued recall, Trail Making Test Part A (TMT-A), Trail Making Test Part B (TMT-B), and Boston Naming Test (BNT)) [39]. No correlation was also found between MBI-C and the sum of Clinical Dementia Rating (CDR) [39]. In accordance, there are several reports suggesting that NPSs in FTD, including apathy, are not directly correlated to cognitive performance, including executive function [43]. On the other hand, the MBI-C was at least weakly correlated with ADL score, supporting the consideration that the daily functioning of patients with bvFTD is influenced at least to some extent by NPS [39], and that the MBI, by definition, is accompanied by at least minimum impaired daily functioning [1].
Concerning genetic FTSD, using Genetic FTD Initiative (GENFI) data, Tavares et al. examined symptomatic mutation carriers in MAPT, GRN and C9orf72, as well as their at-risk relatives to define the earliest clinical features of genetic FTLD [48]. In symptomatic individuals, the most frequently endorsed initial symptoms were apathy and disinhibition, followed by cognitive deficits including memory impairment and decreased fluency [48]. In the preclinical stage, MAPT carriers reported more mood- and sleep-related symptoms, C9orf72 carriers demonstrated a slightly greater degree of abnormal behaviors, while GRN carriers showed fewer mood symptoms compared to non-carriers [48]. Overall, the findings highlight that early behavioral symptoms, particularly the MBI-related domains of decreased motivation and impulse dyscontrol, might be common first manifestations in genetic FTD cases, and may serve as clinically meaningful endpoints for prodromal genetic FTD studies. On the other hand, affective dysregulation domain reflected by mood symptoms might be early, underrecognized signs of prodromal FTD especially in MAPT mutation carriers. In accordance, another GENFI cohort study indicated that the frequency and severity of NPSs differed by genotype (MAPT, GRN, C9orf72) in FTD, showing distinct trajectories across disease stages [49]. MAPT carriers exhibited the highest frequency and severity of several core behavioral features, particularly disinhibition and compulsive behaviors, compared with C9orf72 and GRN carriers [49]. Anxiety and depression were most prominent in GRN and MAPT carriers, whereas hallucinations, including auditory and visual types, were most frequent in C9orf72 carriers [49]. Importantly, psychosis spectrum phenomena appear to carry staging value beyond standard clinical staging scales, supporting their systematic assessment especially in prodromal genetic FTSD [50]. Across the disease course, most symptoms increased in early–intermediate stages and then plateaued, highlighting genotype-specific behavioral signatures that are relevant for phenotyping and trial design. In particular, the disproportionally higher frequency of the MBI domain of psychosis in C9orf72 FTD carriers supports this feature as a potential genotype-enriched prodromal signal rather than a nonspecific late complication. This evidence aligns with the findings by Cui et al. described above, where psychosis was a generally rare manifestation in bvFTD [39]. Thus, systematic capture of psychotic phenomena within an MBI framework may improve early phenotyping and aid suspicion of C9orf72-associated FTSD, particularly when prominent psychosis accompanies subtle behavioral change.
In another study based on a large GENFI cohort of presymptomatic MAPT, GRN, and C9orf72 mutation carriers, followed annually with MRI for two years, apathy increased over time in carriers but not in non-carrier relatives [51]. Importantly, baseline apathy predicted subsequent cognitive decline over the follow-up period, whereas baseline cognition did not predict worsening apathy [51]. Progression of apathy was linked to lower baseline gray-matter volume in frontal and cingulate regions, suggesting a neuroanatomical substrate in motivational control networks [51]. Together, these data position apathy in the context of decreased motivation MBI domain as a potential prodromal neurobehavioral marker in presymptomatic genetic FTD that anticipates subclinical cognitive deterioration.
In summary, current evidence supports MBI as a promising framework for identifying the earliest NPSs of FTSD. By capturing persistent late-life behavioral change across key domains, MBI may improve early recognition, refine phenotyping, and strengthen links between NPSs and underlying neurodegenerative pathology.
4.3. MBI in Four Repeat (4R)-Tauopathies
4R-tauopathies constitute a subgroup of primary tauopathies characterized by the abnormal aggregation of tau isoforms containing four microtubule-binding repeats. Under physiological conditions, alternative splicing of the MAPT gene generates tau isoforms with either 3R or 4R domains. In 4R-tauopathies, this balance is disrupted, leading to preferential accumulation of 4R tau in neurons and glial cells [52]. The main clinicopathological entities in this group include PSP and CBD, although 4R tau pathology may also underly other phenotypes within the FTLD spectrum [52]. Unlike AD, which contains mixed 3R/4R tau and is typically associated with concomitant amyloid pathology, primary 4R tauopathies are defined by distinct molecular, cellular, and anatomical patterns of tau deposition, with prominent involvement of subcortical, brainstem, and frontostriatal networks [53]. These disorders are therefore of particular interest in studies of MBI, because their regional tau burden can be closely linked to syndromes of apathy, dysexecutive dysfunction, behavioral change, and impaired social–emotional regulation.
Dedicated studies examining MBI as a formal syndrome in primary 4R-tauopathies remain very limited, and this represents an important gap in the literature. At present, most relevant evidence comes indirectly from studies of neuropsychiatric and behavioral symptoms in PSP and corticobasal syndrome (CBS) or CBD, rather than from MBI-C-based or ISTAART-AA-anchored investigations.
The available evidence suggests that behavioral changes are clinically relevant early in 4R-taupathies and may be particularly informative in relation to underlying 4R-tau network degeneration. In PSP, apathy is one of the most consistent neuropsychiatric features, with a weighted mean prevalence of approximately 60% across studies, and it may be identifiable even in relatively early disease stages, where it can help distinguish PSP phenotypes from PD [54].
In newly diagnosed PSP and CBS cohorts, apathy, aspontaneity, depression, irritability, and language-behavioral changes have all been described, with apathy more frequent in PSP than CBS (approximately 58% versus 34%) in one study [55]. Moreover, apathy and impulsivity frequently co-occur in PSP, supporting the idea that these syndromes might reflect a disruption of shared frontostriatal and subcortical regulatory systems [56]. This possible interpretation is biologically plausible, as behavioral severity in PSP has been linked to volume loss in the lateral posterior frontal cortex, a region within the frontostriatal networks; meanwhile, apathy specifically has been associated additionally with putaminal atrophy [57], potentially reflecting broader subcortical degeneration.
Complementary 7T MRI data further implicate locus coeruleus (LC) degeneration in apathy and cognition in PSP, suggesting a potential tau-mediated noradrenergic dysfunction in PSP [58]. Interestingly, in comparison to AD and healthy ageing where higher levels of neuronal loss have been shown in the rostral LC, the caudal subregion was mostly affected in PSP in this study [58]. It has been hypothesized that the caudal subregion might be particularly vulnerable to potential environmental toxic stimuli via the CSF due to the proximity to the fourth ventricle [58]. Furthermore, the caudal subregions of the LC receives vagal nerve projections through the solitary tract nucleus, which might exert higher vulnerability to misfolded proteins such as tau protein, transmitted from the periphery [58]. These observations suggest that early NPSs in PSP might arise not only from frontostriatal degeneration, but also from involvement of brainstem neuromodulatory systems, with the LC representing a plausible link between 4R-tau pathology and the emergence of apathy and cognitive dysfunction.
In CBS/CBD, early behavioral presentations appear somewhat less stereotyped, but frontal behavioral symptoms, depression, compulsive features, and irritability are well recognized [59], and early frontal-type behavioral changes may even help predict underlying CBD rather than AD pathology in patients presenting with CBS [60].
Overall, although it would be premature to claim an established MBI phenotype for 4R-tauopathies, the available evidence supports the hypothesis that later-life-emergent apathy, impulse dyscontrol, aspontaneity, and related frontal–subcortical behavioral syndromes may represent clinically meaningful prodromal or early manifestations of PSP and CBD, warranting future studies that apply formal MBI criteria and domain-based phenotyping in these disorders.
6. Potential Therapeutic Candidates Targeting MBI
Given the increasing amount of evidence supporting the association between MBI and dementia, it can be hypothesized that treatment strategies targeting mild NPSs in subjects with normal cognition or MCI would be effective, particularly in reducing the risk of developing subsequent cognitive decline. However, future studies are necessary to determine whether MBI is a reversible condition [6]. Although no approved drug exists that specifically targets MBI, numerous pharmacological compounds as well as non-pharmacological approaches have shown potential benefits.
Among them, citalopram, a selective serotonin reuptake inhibitor (SSRI), has been reported to exhibit neuroprotective effects by promoting neurogenesis, suppressing neuroinflammation, and lowering amyloid and tau pathology [110]. SSRI treatment in patients with MCI and depression has also been associated with delayed progression to AD [111]. The use of cholinesterase inhibitors, specifically donepezil, has improved both NPSs and cognitive symptoms, particularly in patients with severe AD [112,113,114]. Noradrenergic and norepinephrine blocking agents, like prazosin and methylphenidate, were also found to be effective in alleviating apathy, agitation, and aggressiveness [115,116]. Moreover, other pharmacological compounds were explored showing promising results. In particular, phosphodiesterase-4 (PDE4) inhibitors, such as rolipram and roflumilast, have been shown to improve both cognitive function and depression in animal models, possibly by modulating the hypothalamic–pituitary–adrenal (HPA) axis, the cyclic AMP response-element-binding protein (CREB)/BDNF signaling pathway as well as anti-inflammatory mechanisms [117,118,119,120]. Additional agents, including melatonin, apelin-13, nattokinase, the antibiotic minocycline, and the ATP-sensitive potassium-channel inhibitor glibenclamide, reported a promising therapeutic potential, particularly in the domains of affective dysregulation and apathy in AD, through the regulation of HPA axis and neuroinflammatory responses [121,122,123,124,125]. Combined therapies, including antidepressant and acetylcholinesterase inhibitors, may further improve clinical outcomes, especially when mood disorders and cognitive impairment are diagnosed [126].
Notably, non-pharmacological interventions, like physical exercise and swimming, have also been linked to a lower risk of dementia by decreasing tau hyperphosphorylation, reducing amyloid deposition, preserving synaptic density, as well as suppressing neuroinflammation and neuronal damage [127]. Furthermore, these approaches have demonstrated promising results in ameliorating both cognitive deficiencies and behavioral symptoms, such as anxiety and depression [128,129,130].
7. Future Directions and Translational Implications
Despite the growing promise of MBI as an early behavioral framework across tauopathies, several conceptual and methodological challenges remain. While the MBI-C has brought standardization, the retrospective application of the MBI construct in previous datasets, such as using NPI data, has several limitations, including the shorter time frame, and the specificity of NPI mainly for dementia. As a result, symptom duration thresholds, domain boundaries, and cross-study comparability remain imperfect. In frontotemporal and other non-AD tauopathy cohorts, the optimal MBI-C cutoffs for case identification, prognostic enrichment, and differential diagnosis also remain insufficiently established.
Importantly, recent works highlight that scalable blood-based plasma biomarkers may support the study of MBI across the continuum of neurodegenerative diseases. In the early stages, diagnostic biomarkers may potentially help discriminate NPSs in the context of neurodegeneration from primary psychiatric syndromes and identifying the underlying neuropathology, whereas in dementia-stage disease, biomarkers might help track symptom trajectories and treatment responses [5,131]. In this context, growing evidence links MBI status and/or specific MBI domains with plasma biomarkers across the cognitive continuum, including associations with lower plasma Aβ42/40, longitudinal increases in plasma p-tau181, and higher plasma p-tau217 in dementia-free older adults [5,131]. These findings support the hypothesis that the combined use of MBI and plasma biomarkers might help in the differential diagnosis of the underlying neuropathology, especially in discriminating between AD and non-AD diseases. From a practical standpoint, MBI screening could serve as a low-cost triage step to select dementia-free individuals for confirmatory AD biomarker testing, such as p-tau217, potentially improving efficiency in prevention-oriented cohorts and trials [132,133].
The differentiation of MBI from primary psychiatric conditions remains difficult, particularly in younger patients, thereby resulting in potential diagnostic misclassification. Novel biomarkers, such as NfL may help towards this direction. In this regard, plasma NfL has shown high discriminatory performance in differentiating bvFTD from primary psychiatric disorders, such as bipolar disorder, major depression, or treatment-resistant schizophrenia [134]. In another study, serum NfL also differentiated bvFTD from primary psychiatric disorders with very good sensitivity and specificity, supporting the robustness of NfL across settings and cohorts [135]. This is highly relevant to the MBI framework, as many bvFTD presentations initially meet MBI criteria before overt dementia. Hence, combining MBI-based phenotyping with NfL could help flag neurodegeneration in late-onset psychiatric presentations that otherwise risk diagnostic delay. For astroglial markers, evidence suggests GFAP differs across etiologies and is not as strong as NfL for distinguishing bvFTD from primary psychiatric disorders. Plasma GFAP has been shown to be higher in bvFTD, but still inferior to NfL [134]. In broader FTD biomarker work, plasma GFAP has been found elevated in FTD versus cognitively normal controls, but lower than in AD [136]. Higher baseline GFAP has also been associated with faster longitudinal cognitive decline and increased risk of progression to cognitive impairment in FTD [136]. Finally, regarding tau-related markers, plasma total tau has been reported to be higher in clinical FTD syndromes, including bvFTD, versus healthy controls, but shows genetic subtype effects, with significant increases in symptomatic MAPT mutation carriers, and not consistently in C9orf72 or GRN [137]. Taken together, these findings support a biomarker-informed approach in which MBI-based behavioral phenotyping, particularly when combined with plasma NfL and selected glial or tau-related markers, may improve the early distinction between prodromal neurodegeneration and primary psychiatric disease.
Furthermore, despite behavioral symptoms being common, many neurodegenerative disease trials still exclude participants with significant psychiatric symptoms, paradoxically omitting those most relevant to the MBI construct. While predictive associations for MBI symptoms are relatively strong in AD-related cohorts, large-scale, prospective studies in other tauopathies are limited. These limitations reinforce the need for refined instruments and cutoff points, more uniform stratification methods, and more inclusive research designs.
Several priorities should shape the next phase of translational research. First, prospective longitudinal studies are needed to validate MBI domains against multimodal biomarkers, including plasma and CSF NfL, tau species and isoform-sensitive assays, inflammatory markers, and structural and functional network imaging. Such studies should move beyond total MBI burden and examine whether specific domain profiles map onto distinct molecular signatures, patterns of network disintegration, and rates of progression. Second, genotype-informed phenotyping may help define mutation-specific prodromal behavioral trajectories and clarify whether MBI can serve as a trans-syndromic but biologically informative marker of underlying pathology. Third, domain-level neuroimaging approaches should examine whether decreased motivation, affective dysregulation, impulse dyscontrol, and social inappropriateness can be reliably linked to selective disruption of frontostriatal, salience, default mode, or socioemotional networks before overt dementia emerges. Fourth, digital phenotyping offers a major opportunity for scale: passive behavioral monitoring, speech and language analysis, actigraphy, social interaction metrics, facial affect recognition, and ecologically sampled behavioral data may help operationalize MBI continuously and sensitively outside the clinic. Finally, MBI should be incorporated into experimental medicine and early-phase intervention trials, not only as a screening or enrichment tool, but also as a potential outcome measure for therapies targeting synaptic dysfunction, network instability, neuroinflammation, or pathogenic protein spread. In this context, MBI may ultimately contribute to a more biologically grounded stratification of neurodegenerative disease, refine subtype classification within existing diagnostic categories, and support earlier, more personalized behavioral and disease-modifying interventions.
8. Conclusions—Reframing Behavior in Neurodegeneration
MBI offers a clinically meaningful framework for reinterpreting later-life-emergent NPSs as potential early manifestations of neurodegenerative disease. Across the spectrum of tauopathies, converging clinical, neuroimaging, molecular, and biomarker evidence suggests that persistent behavioral change may reflect the selective vulnerability of frontolimbic, salience, default mode, and frontostriatal networks to tau-mediated dysfunction. In this context, MBI domains can be viewed as a useful marker of early circuit failure, linking symptoms such as apathy, affective dysregulation, impulsivity, disinhibition, and altered social conduct to underlying synaptic, neuromodulatory, inflammatory, and network-level pathology. This perspective is particularly valuable in disorders such as AD, bvFTD, PSP, and CBD, where behavioral symptoms may precede overt dementia or major motor syndromes and therefore provide an important window into prodromal disease.
At a translational level, MBI has the potential to serve as a link between bedside phenotyping and biomarker-based neurodegeneration research. When integrated with fluid biomarkers, genotype-informed stratification, and multimodal neuroimaging, MBI may improve early detection, refine differential diagnosis, enrich clinical trial populations, and offer domain-sensitive outcome measures for prevention-oriented and disease-modifying interventions. Reframing behavior in this way shifts NPSs from the periphery to the center of neurodegenerative disease models. Recognizing and operationalizing MBI as the possibly earliest clinically accessible expression of tau-related brain dysfunction may help towards advancing earlier, biologically informed, and more personalized approaches to the spectrum of tauopathies.
Acknowledgments
All figures were created using BioRender (Villa, C. (2026) https://BioRender.com/fbqa07u, accessed on 11 March 2026; https://BioRender.com/4llgthn, accessed on 30 March 2026; https://BioRender.com/5rr28vy, accessed on 30 March 2026; https://BioRender.com/468n092, accessed on 30 March 2026.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
No new data were created or analyzed in this study.
Conflicts of Interest
The authors declare no conflicts of interest.
Abbreviations
The following abbreviations are used in this manuscript: ADAlzheimer’s diseaseALSamyotrophic lateral sclerosisaMCIamnestic MCIAVLTauditory verbal learning testBNTBoston naming testBPSDbehavioral and psychological symptoms of dementiabvFTDbehavioral variant FTDCBDcorticobasal degenerationCBScorticobasal syndromeCDRClinical Dementia RatingCREBcyclic AMP response-element-binding proteinCSFcerebrospinal fluidDATdopamine transporterDRNdorsal raphe nucleusFBIFrontal Behavioral InventoryFTDfrontotemporal dementiaFTLDfrontotemporal lobar degenerationFTLD-taufrontotemporal lobar degeneration with tau pathologyFTSDfrontotemporal spectrum disordersGDS-15geriatric depression scale-15GENFIgenetic FTD initiativeHPAhypothalamic–pituitary–adrenalHThydroxytryptamineILinterleukinISTAART-AAinternational society to advance Alzheimer’s research and treatment-Alzheimer’s associationLClocus coeruleusMBImild behavioral impairmentMBI-CMBI checklistMCImild cognitive impairmentMNDsmotor neuron diseasesNF-κBnuclear factor kappa-light-chain-enhancer of activated B cellsNfLneurofilament lightNFTsneurofibrillary tanglesNPINeuropsychiatric InventoryNPI-QNPI-questionnaireNPSneuropsychiatric symptomsPDParkinson’s diseasePDE4phosphodiesterase-4PETpositron emission tomographyPIAprofessional interest areaPSPsupranuclear palsySCDsubjective cognitive declineSERTserotonin reuptake transporterSSRIselective serotonin reuptake inhibitorSV2Asynaptic vesicle glycoprotein 2ATMT-Atrail making test part ATMT-Btrail making test part BTREM2triggering receptor expressed on myeloid cells 2VAChTvesicular acetylcholine transporter