• About DELAD

    DELAD is an international initiative dedicated to the collection, organisation, archiving, and sharing of language and speech disorders corpora. DELAD supports the scientific community by:

    • Providing access to high-quality datasets
    • Promoting best practices in data management and ethical use
    • Supporting the long-term preservation of sensitive research materials

    DELAD facilitates collaboration between researchers, infrastructure providers, and practitioners working in clinical linguistics, speech-language pathology, phonetics, and related disciplines

    Learn more about DElad

    In association with

    Keynote Speaker

    Language Disorders 2026 Online

    Iris Edda Nowenstein

    University of Iceland

    Main areas of research

    • First language acquisition and multilingualism
    • The memory-language interface
    • Speech and language disorders
    • Clinical linguistics
    • Clinical language technology
    • Case marking and argument structure
    • Language variation and change
    Publications

    Bio

    Iris Nowenstein is an Assistant Professor in Icelandic Linguistics at the University of Iceland and a Speech-Language Pathologist at the National Universal Hospital of Iceland. She defended her PhD dissertation on the acquisition of Icelandic case marking at University of Iceland in 2023 and was a Fulbright Visiting Scholar at the University of Pennsylvania in 2024-2025, within the Integrated Language Science and Technology initiative. Her current work is within clinical linguistics and clinical language technology, focusing on the automatic extraction of language-specific and language-universal manifestations of diseases and disorders. She is involved in various efforts to ensure that people in smaller and lesser-resourced language communities have access to digital health resources which are based on language technology.

    Keynote Lecture

    Language and Speech Disorders Corpora as Necessary Infrastructure for Health Tech Accessibility

    Recent clinical applications of language technology show considerable potential for people with speech and language symptoms and disorders. This has resulted in a variety of health tech tools becoming available to them, including novel communication aids such as personalized automatic speech recognition for disordered speech and novel instruments for the screening, diagnosis and monitoring of diseases and disorders via automatic language sample analysis. This kind of monitoring through speech might, for example, provide cost-effective, person-centered and non-invasive endpoints for treatment efficacy assessments, including in the context of precision medicine and drug trials.

    But crucially, these speech- and language-based health tech tools are currently almost exclusively accessible to speakers of English and a few other high-resource languages. A major hurdle for small, lower-resourced language communities in this context is the creation of clinical corpora, including language and speech disorders corpora. These corpora often constitute the basis for the successful generalization of technology across languages, both in the context of diagnosis and monitoring (where manifestations of diseases and disorders can be language-specific) and communication aids, where accuracy gains can heavily rely on dataset size and quality. 

    In this talk, I will describe ongoing efforts to build the necessary infrastructure for clinical speech and language data collection in Iceland. This is done through the creation of the Icelandic Language Biobank, a resource that leverages collaboration with clinicians and robust linguistically-informed data collection against data scarcity. 

    In contrast with the most comprehensive data collection efforts in high-resource languages, the Icelandic Language Biobank will contain language samples for communication aid development as well as the diagnosis and monitoring of diseases and disorders. Data will primarily be collected through a web-based automatic linguistic analysis platform, ALDA (Automatic Linguistic Data Analysis), designed for and co-created with speech-language pathologists/therapists (SLPs). A platform which combines accessible clinician-centered tools and a data sharing infrastructure serves the purpose of facilitating technological transfer and  creating conditions for sustainable clinician-led data collection. I argue that taking these steps brings us closer to the language and speech disorders corpora that are necessary to ensure access to speech- and language-based health tech tools in small, less-resourced language communities.

    • Data stewards interested in sensitive language data
  • Language Disorders 2026 Online

    Stefano F. Cappa

    IUSS School for Advanced Studies, Pavia & IRCCS Istituto Auxologico, Milan, Italy

    Stefano Cappa

    Main areas of research

    • Language Disorders and Aphasia
    • Alzheimer’s Disease and Frontotemporal Dementia
    • Digital Cognitive Assessment
    • Brain–Language Relationships
    • Semantic Memory and Meaning Representation
    Publications

    Bio

    Stefano F. Cappa is a neurologist and cognitive neuroscientist. Professor Emeritus of Neurology at the University Institute for Advanced Studies (IUSS), he is presently a consulting neurologist at the IRCCS Istituto Auxologico in Milan. His main areas of clinical and research activity are disorders of language, memory, and social cognition, with a focus on Alzheimer’s disease and frontotemporal dementia. His main scientific contributions deal with the representation of meaning in the brain (semantic memory), its degradation due to brain disorders, and the mechanisms of recovery. In recent years, his interests have extended to brain health, the prevention of memory decline, and the development of digital assessments. He has published more than 400 papers in peer-reviewed journals and several books on aphasia and cognitive neurology.

    Research Interests

    • Linguistic aspects of aphasia, with particular emphasis on the neurological correlates of language impairment
    • Alzheimer’s disease and other dementias
    • Diagnosis and neuropsychological features of memory disorders in dementia
    • Application of functional imaging methods (positron emission tomography and functional magnetic resonance imaging) to the study of cognitive functions, particularly language and semantic memory
    • Cognitive studies of spatial cognition and exploration disorders, their neurological correlates, and factors influencing recovery
    Keynote Lecture

    The contribution of Primary Progressiva Aphasia to language neurobiology

    The primary progressive aphasias (PPA) are clinical syndromes characterised by progressive language dysfunction, reflecting an (initially) selective degeneration of the brain language networks due to a variety of neuropathological mechanisms.

    The enormous research interest in PPA, testified by the impressive increase in number of publications during the last decade, can be attributed to several reasons:

    – the detailed analysis of spared/impaired aspects of language dysfunction in PPA provides a model of progressive and selective impairment, complementary to the evidence coming from aphasia due to stroke, which, in contrast, is characterised by the sudden onset of language impairment and by possible recovery

    -at the neurological level, the PPA syndromes are characterised, at least in the early to moderate stages, by selective patterns of involvement of grey and white matter, which are generally consistent at the single subject level, and can be explored in vivo by multiple structural and functional imaging methods, allowing insights into brain segregation and integration of the neurological mechanisms of language function The study of PPA syndromes is providing  important insights into multiple aspects of language neurobiology. In particular, patients living with the semantic variant have crucially contributed to the understanding of the brain organization of semantic knowledge.  The non fluent/agrammatic variant offers a window into the brain mechanism of speech production and syntactic processing. The logopenic/phonological variant  is specially informative about the relation of other cognitive functions, such as short-term memory,  with language performance. Recent development in this field of study are based on the promising application of AI methods to the important issue of rehabilitation and to the development of cross-linguistic investigations.

  • Language Disorders 2026 Online

    Bonnie Nozari

    Carnegie Mellon University

    Bonnie Nozari

    Main areas of research

    • Psycholinguistics
    • Cognitive psychology
    • Cognitive neuroscience
    • Aphasiology
    • Computational modelling of language processes
    Publications

    Bio

    Bonnie Nozari is a professor of psychology and cognitive science at Indiana University (IU). Before joining IU, she was a faculty at Johns Hopkins’ Neurology Department, and later at Carnegie Mellon University’s Department of Psychology. Her research focuses on understanding the computational and neural basis of language production. An important part of this investigation is studies of aphasia, i.e., language impairment after brain damage, as such studies help us better understand neurotypical systems, and also find solutions for restoring the lost linguistic functions in clinical populations.

    Research Interests

    • Language production and the cognitive mechanisms underlying the transformation of thought into speech
    • Efficiency and control in complex language production systems
    • Interactions between language production, monitoring, and executive control
    • Domain-generality of cognitive control mechanisms across language and other cognitive systems
    • Language production across populations, including children and stroke survivors
    • Developmental change and recovery-related reorganisation in language systems
    Keynote Lecture

    Leveraging big data and AI in studies and treatment of aphasia

    Speech errors have been instrumental in advancing our understanding of the architecture of the language production system, the nature of its representations, and its disorders. To be most informative, researchers usually need large amounts of data. Hand-coding such data can be both cumbersome and subjective. In this talk, I will present LeCoder, the first open-source, automated error coder, which uses a data-driven approach grounded in large-scale corpora to quantify the target-response relationship, allowing it to be flexible, scalable, and generalizable across new datasets.

    By testing the coder on two datasets from two aphasia labs that have been carefully coded by trained research assistants, I will show that LeCoder has high accuracy when compared to expert coders, and in certain cases, offers a more logical categorization than human coders. I will then show, using robust machine-learning approaches, that LeCoder’s performance generalizes to new participants and items it has never encountered before. Collectively, these findings encourage the use of LeCoder across labs for more objective coding of speech errors, which will, in turn, increase replicability of findings in all subfields of research that use speech error analysis, including neuropsychological research.

    Time permitting, I will also briefly preview a new platform developed in my lab, which uses large language models to help individuals with agrammatical aphasia produce more complete sentences in real time.