Linguistic Creativity and Formal Thought Disorder in Schizophrenia
Oli’s PhD looked at how linguistic creativity (poetry, song lyrics, figurative language) might be related to formal thought disorder in schizophrenia. Formal thought disorder (FTD) refers to a group of transdiagnostic language symptoms that, in schizophrenia, can be viewed in terms of positive, negative, and disorganisation syndromes.
The results showed that language disturbances are on a continuum with normal speech, with nonpsychiatric speakers producing the same types of errors as clinical speakers but at lower frequencies. The results also showed that clinical speakers have more difficulty maintaining the conversational topic, tending to over-involve issues of personal and emotional concern.
In schizophrenia, the comprehension of linguistic creativity is adversely affected. There is not enough evidence to say whether the production of linguistic creativity is affected similarly, although early evidence suggests that it might be spared. The comprehension of linguistic creativity is more likely to be preserved in the affective psychoses, particularly in cases where positive symptoms are the dominant problem.
These findings are consistent with those of other FTD researchers, such as Liddle and Palaniyappan, and Oli is currently collaborating with these experts on future projects.
The Discussing Abstract Ideas in Schizophrenia Corpus (DAIS-C)
The DAIS-C is a speech corpus. It is composed of roughly twenty hours of unplanned speech, produced by 15 speakers with a formal diagnosis of schizophrenia and 14 speakers with no self-reported psychiatric history. The corpus was built according to Sinclair’s corpus construction guidelines, and it is currently the only corpus of its type available to researchers for free with no restrictions.
Data were collected using an unstructured interviewing approach. This makes the dataset unsuitable for qualitative content analysis. It is however suited to quantitative and qualitative analyses that focus on computational and interactional questions, such as training NLP systems and assessing pragmatics.
The Deciding on Literary Imagery (DOLI) Task
The DOLI Task is a psycholinguistic experiment. It combines elements of the Pyramid and Palm Trees Test (Howard and Patterson, 1992), minimal context metaphor production tasks (Glucksberg et al., 1997), and Shen’s CCT tasks (Shen and Aisenmann, 2008). Participants view primes composed of adjective pairs (which act as the source domain) and make selections between concrete and abstract nouns (which constitute targets). Task engagement creates a joint-modified synaesthetic metaphor, a generally understudied metaphor type.
Researchers can use the DOLI task to test whether synaesthesia drives the creation of content similarity-based (concrete choice) or domain similarity-based metaphors (abstract choice). Its design also allows researchers to study the role of prime type, the order of the noun choices, and more. It can be applied to clinical and nonclinical populations, as in as in Delgaram-Nejad et al. (2022) and Shen and Aisenman (2008) respectively.
The task was intially computerised using Superlab (Cedrus Corporation) and delivered, during the COVID pandemic, with Superlab Remote. Licenses for these programs are required to use the computerised files, which have unfortunately since been lost, but the raw stimuli lists are available in my thesis (see above). I recommend that Python is used to deliver the experiment in the future. I am currently working on such a version in my spare time. Please contact me to discuss using the DOLI stimuli in your own work.