VRT (verbal reasoning test): a new test for assessment of verbal reasoning. Test realization and Italian normative data from a multicentric study

Challenging large language models' "intelligence" with human tools: A neuropsychological investigation in Italian language on prefrontal functioning

The Artificial Intelligence (AI) research community has used ad-hoc benchmarks to measure the "intelligence" level of Large Language Models (LLMs). In humans, intelligence is closely linked to the functional integrity of the prefrontal lobes, which are essential for higher-order cognitive processes. Previous research has found that LLMs struggle with cognitive tasks that rely on these prefrontal functions, highlighting a significant challenge in replicating human-like intelligence. In December 2022, OpenAI released ChatGPT, a new chatbot based on the GPT-3.5 model that quickly gained popularity for its impressive ability to understand and respond to human instructions, suggesting a significant step towards intelligent behaviour in AI. Therefore, to rigorously investigate LLMs' level of "intelligence," we evaluated the GPT-3.5 and GPT-4 versions through a neuropsychological assessment using tests in the Italian language routinely employed to assess prefrontal functioning in humans. The same tests were also administered to Claude2 and Llama2 to verify whether similar language models perform similarly in prefrontal tests. When using human performance as a reference, GPT-3.5 showed inhomogeneous results on prefrontal tests, with some tests well above average, others in the lower range, and others frankly impaired. Specifically, we have identified poor planning abilities and difficulty in recognising semantic absurdities and understanding others' intentions and mental states. Claude2 exhibited a similar pattern to GPT-3.5, while Llama2 performed poorly in almost all tests. These inconsistent profiles highlight how LLMs' emergent abilities do not yet mimic human cognitive functioning. The sole exception was GPT-4, which performed within the normative range for all the tasks except planning. Furthermore, we showed how standardised neuropsychological batteries developed to assess human cognitive functions may be suitable for challenging LLMs' performance.

Cognitive profile in a young woman with Susac syndrome: a case report

Susac syndrome is a rare immune-mediated endotheliopathy affecting the microvasculature. It presents three main symptoms: encephalopathy, branch retinal artery occlusions, and hearing loss . Here we present a new case report focusing on the evolution of the cognitive profile. The patient underwent two neuropsychological examinations. The first, one month after the onset of the symptomatology, highlighted a prevalent involvement of verbal executive functions. The second, conducted six months later, revealed a global improvement in most previously deficient areas, although with the persistence of a difficulty in cognitive estimation. . This case illustrates the importance of a comprehensive analysis of patients with Susac syndrome to appreciate the whole range of cognitive deficits and reliably evaluate symptom evolution.

Verbal Reasoning Impairment in Parkinson’s Disease

Background. The aim of this study was to assess verbal reasoning (VR) functioning in patients with Parkinson’s disease (PD) and healthy controls (HCs). Methods. The non-demented PD patients and HCs matched by age and global cognition were enrolled in this study. VR was assessed with the verbal reasoning test (VRT), total score, and subsets. Results. Eighty-seven PD patients (51 men; mean age $63.8\pm 7.9$ years) and 87 HCs (46 men; mean age $63.7\pm 8.0$ years) were enrolled. At univariate analysis, PD patients presented a significantly lower score in the VRT subset classification ( $12.3\pm 2.1$ ) than HCs ( $12.9\pm 1.7$ ) with an odds ratio (OR) of 0.8 (95% confidence interval [CI] 0.70–0.98; $p=0.003$ ). The strength of association was also confirmed at multivariate analysis (OR = 0.8, 95% CI 0.70–0.98; $p=0.003$ ). Moreover, in PD patients, a statistically significant positive correlation was found between VRT-classification and MoCA scores ( $r=0.330$ ; $p=0.002$ ). Conclusions. PD patients presented lower VR performance than HCs.

Normative Values of the Groffman Visual Tracing Test for the Assessment of Oculomotor Performance in the Adult Population

The Groffman visual tracing (GVT) test is an indirect psychometric measure of oculomotor performance, used for the clinical assessment of eye movements. The test consists of two cards with five contorted lines of increasing overlap, crowding, and difficulty. The task starts from each of the letters at the top of the page, follows the line from the letter to the corresponding number at the bottom of the page, and the number is named. Although the GVT test was developed for the evaluation of children, it has also been applied to adults with visual and cognitive deficits. However, it lacks reference values. Therefore, the aim of the study was to assess oculomotor behavior across the typical human lifespan and to define normative data in an adult population. A total of 526 adults aged between 20 and 79 years, all without neurological or psychiatric deficits, were enrolled in the study. The results were analyzed by considering the accuracy and execution times separately. An influence of age, education and sex for accuracy was found, and age for the execution times was found. Norms for adults were developed considering the specific structure of the test and the accuracy and the execution time separately. The GVT test can now be applied in healthy and neurological adult populations for the evaluation of oculomotor performance.

On the (lack of) association between theory of mind and executive functions: a study in a non-clinical adult sample

We investigated in a sample of non-clinical adults the association between Theory of Mind (ToM) and Executive Functions (EFs), that is the set of skills that allow people to control and modulate lower-level cognitive processes in order to produce appropriate behaviour. To this aim, we assessed both affective (i.e., understanding other people's emotions and feelings) and cognitive (i.e., understanding others' beliefs and intentions) ToM, as well four subcomponents of EFs, that is Updating, Shifting, Inhibition, and Access. The association between ToM and non-verbal fluid intelligence, verbal reasoning, and cognitive estimation abilities was also investigated. Eighty-one healthy participants were recruited, and a set of psychometrically validated tests was administered. Multiple regression analyses were run to assess significant predictors of ToM performance when potentially confounding predictors (sociodemographic variables) were controlled for. Results showed a lack of association between affective/cognitive ToM and EFs, whereas non-verbal fluid intelligence for cognitive ToM and verbal reasoning for affective ToM were found to be significant predictors of ToM performance. These results represent a contribution toward a deeper understanding of the ToM-EFs relationships and highlight the importance of broadening the analysis of these relationships to the role played by other domain-general functions in both affective and cognitive ToM.

The Verbal Judgement Task: Normative data of verbal abstract reasoning in a sample of 18- to 40-years old

In this study, normative data for the age-range 18-40 years have been provided for the Verbal Judgment Test (VJT), which underpins abstract reasoning on the basis of four subtests: "Differences", "Proverbs", "Absurdities" and "Classifications". 554 participants (280 males and 274 females) were recruited and the following data were provided: means and standard deviations divided by gender, educational level (8, 13 and 18 years) and age group (18-20 years, 21-25 years, 26-30 years, 31-35 years and 36-40 years) for each subtest and the total score; percentiles for each subtest, divided by age group, and, when appropriate, educational level and/or gender; Rho correlations between age group, gender, educational level, intelligence and VJT scores. Age-, education- and gender differences were also assessed carrying out non parametric tests. Results showed that age and education positively affected performance in the subtests of Differences, Proverbs and Classifications, which are mostly based on previous knowledge, experience, and crystallized intelligence, but did not affect performance in the Absurdities subtest, which encompasses to some extent fluid intelligence. In addition, males showed higher scores than females in the subtests of Differences and Proverbs and in the total VJT, probably reflecting higher knowledge acquisition. Implications for future research are briefly discussed.