The right anterior temporal lobe has emerged over the past decade as a hub of what researchers call socioemotional semantics: the stored knowledge that allows us to understand what emotions are, what they mean, and how they apply to particular people and situations. Whereas the left temporal lobe is dominant for verbal and factual semantic knowledge, the right hemisphere appears to carry a disproportionate share of conceptual knowledge about the social and emotional world. Work by Younes and colleagues, published in Brain in 2022, argued that right temporal degeneration produces a distinctive semantic variant of behavioral frontotemporal dementia in which this socioemotional knowledge base erodes. Patients may retain vocabulary and general facts while losing the ability to grasp the emotional significance of faces, voices, and situations. The consensus statement from the International Working Group on right temporal predominant frontotemporal dementia, published in Communications Medicine in 2025, formalized this clinical picture and highlighted how frequently the condition goes unrecognized in ordinary practice.

That recognition problem is not trivial. A multicenter retrospective cohort study led by Ulugut and colleagues in Alzheimer’s & Dementia in 2024, spanning dozens of memory clinics across Europe and North America, documented how often patients with right anterior temporal predominance receive delayed or incorrect diagnoses. Because their symptoms are emotional and behavioral rather than linguistic, they are frequently mislabeled as having primary psychiatric disorders, and their brain scans may be misread because radiologists are more accustomed to left-sided temporal atrophy patterns. The clinical stakes of understanding exactly what goes wrong in the emotional brain in this variant are therefore high, both for diagnosis and for counseling families about what to expect as the disease progresses.

The Leuven-led study approaches the problem with a methodological twist. Rather than asking simply which brain regions light up when patients view emotional material, the researchers apply pattern expression analysis, a technique refined by Tor D. Wager and collaborators over the past decade. Pattern expression methods rely on multivariate neural signatures: spatial maps of voxel-level activity weights that have been derived in large independent samples and validated as sensitive and specific markers of particular mental states. Among the signatures relevant to this work are a neural signature for picture-induced negative affect developed by Chang and colleagues in PLoS Biology, a distributed signature for the subjective experience of fear described by Zhou and colleagues in Nature Communications, and dissociable signatures of empathic care and empathic distress described by Ashar and colleagues in Neuron. By projecting each patient’s brain activity patterns onto these signatures, the researchers can estimate, on a continuous scale, how strongly a given emotional process is expressed in an individual brain.

This approach matters because univariate activation analysis, the traditional workhorse of functional MRI, is poorly suited to degenerative diseases. Atrophy and signal loss in the damaged tissue can mask genuine changes in how remaining neural circuitry functions, and emotions are represented in distributed, overlapping networks rather than in single blobs of activation. Pattern expression sidesteps some of these limitations by summarizing whole-network geometry. The technical pipeline underpinning the study draws on well-established neuroimaging infrastructure: structural images processed with FreeSurfer and the Desikan gyral parcellation for regional volume estimates, cortical thickness compared against normative data compiled by Potvin and colleagues, and functional images preprocessed with the FMRIPrep pipeline described by Esteban and colleagues in Nature Methods. These choices place the findings on a reproducible, transparent footing that other groups can adopt.

The theoretical backdrop is the idea, articulated most influentially by Mesulam, that large-scale neurocognitive networks encode knowledge through representation, inference, and transcendent encoding across heterogeneous cortical hubs. Within that framework, the right anterior temporal lobe functions as a convergence zone for nonverbal conceptual knowledge, including knowledge of emotions, faces, and social scripts. Hurley and colleagues demonstrated a nonverbal route to conceptual knowledge involving the right anterior temporal lobe, and Lambon Ralph, Jefferies, Patterson, and Rogers provided the computational account of semantic cognition that situates such hubs within a controlled semantic retrieval system. When the hub degenerates, the downstream consequence is not simply local dysfunction but a degradation of the distributed network’s ability to construct emotional meaning from sensory input.

Previous imaging work in frontotemporal degeneration has already hinted at how this cascade unfolds. De Winter and colleagues showed in Cortex that amygdala atrophy alters emotion-related activity in face-responsive cortical regions, demonstrating that structural damage to one node of the emotion network propagates functionally to distant nodes. Marshall and colleagues mapped the functional neuroanatomy of emotion processing across the frontotemporal dementias in Brain, and Lindberg and colleagues documented altered empathy processing in the disease in JAMA Network Open. Rouse, Binney, Patterson, Rowe, and Lambon Ralph proposed a neuroanatomical and cognitive model of impaired social behavior that integrates atrophy in temporal and frontal networks. The new letter extends this line of inquiry by asking not merely where emotional processing falters but how the characteristic spatial patterns of emotional brain activity are expressed, or fail to be expressed, in right temporal variant patients.

The study was conducted in compliance with the Declaration of Helsinki and approved by the ethics committee of University Hospitals Leuven, with written informed consent from all participants, and the authors report no conflicts of interest. Funding for the work was provided in part by the Research Foundation Flanders, KU Leuven, the King Baudouin Foundation, and the Sequoia Fund for Research on Ageing and Mental Health. The data supporting the findings are available on request from the corresponding author but are not publicly accessible, because they contain information that could compromise participant privacy. Corresponding author Jan Van den Stock, whose laboratory has long focused on social and emotional processing in dementia, led the collaboration together with senior co-authors including Mathieu Vandenbulcke and Dartmouth’s Tor D. Wager.

For the field, the significance of this work lies in its convergence of two research programs that have developed largely in parallel: the clinical nosology of right temporal variant frontotemporal dementia, now consolidated by international consensus, and the computational neuroscience of emotion signatures, now mature enough to be applied to individual patients. If validated patterns of emotional brain response can be measured reliably in this patient group, the same metrics could eventually serve as biomarkers for early diagnosis, as targets for interventions aimed at preserving socioemotional competence, and as outcome measures in trials of disease-modifying therapies. For families coping with a disorder that erodes the emotional core of personhood, the study offers something more modest but no less valuable: a rigorous, mechanistic account of what is happening inside the brain when a loved one’s emotional world begins to change, and a demonstration that modern imaging science can now see that change with increasing precision.

Subject of Research: Emotional brain responses and pattern expressions in right temporal variant frontotemporal dementia

Article Title: Emotional brain responses and pattern expressions in right temporal variant frontotemporal dementia

Article References: Emotional brain responses and pattern expressions in right temporal variant frontotemporal dementia. (n.d.). https://doi.org/10.1007/s00415-026-14117-0

Image Credits: AI Generated

DOI: 10.1007/s00415-026-14117-0

Keywords: frontotemporal dementia, right temporal variant, emotional processing, fMRI, pattern expression, neural signatures, amygdala, socioemotional semantics, right anterior temporal lobe, empathy, neurodegeneration, Journal of Neurology