Delirium is a condition which impacts nearly half of older adults during hospital admission. It presents with a wide range of neuropsychiatric symptoms leading to increased morbidity and mortality. Despite this, specialised knowledge and ownership of the condition remain unclear.

To compare evidence surrounding the roles of neuronal and non-neuronal cells in the overall pathophysiology of delirium and consider the impact this could have in practice.

Using PRISMA systematic review guidelines, five medical research databases were screened for papers discussing the role of neuronal and/or non-neuronal cells in the pathophysiology of delirium between 2011 and 2021.

Fifteen papers which met the inclusion criteria were then categorised into discussing neuronal (n=2), non-neuronal (n=4) or both (n=9) types of cells’ roles in the pathophysiology of delirium. Delirium was often caused by a homeostatic imbalance secondary to acute illness leading to deterioration of neural synapses and therefore signal transmission. However, it was also argued that activated non-neuronal cells, particularly microglia and astrocytes, played a significant role through disruption of the blood brain barrier. This was likely to play a role in the more severe clinical presentations of delirium.

The pathophysiology of delirium is multifactorial with neuronal and non-neuronal cells implicated in neurological disruption. There is no clear agreement on how these mechanisms vary according to aetiology and, ultimately, the severity of delirium. Further research will help refine these theories, which will support the pharmacological and clinical management of the condition.

No significant relationships.

Meta-analyses report moderate effects across cognitive remediation (CR) trials in schizophrenia. However, individual responses are variable, with some participants showing no appreciable gain in cognitive performance. Furthermore, reasons for heterogeneous outcome are undetermined. We examine the extent to which CR outcome is attributable to near learning—direct gains in trained cognitive tasks—while also exploring factors influencing far transfer of gains during training to external cognitive measures.

Thirty-seven schizophrenia outpatients were classified as CR responders and non-responders according to change in MATRICS Consensus Cognitive Battery composite score following 20 sessions of computer-based training. Metrics of near learning during training, as well as baseline demographic, clinical, cognitive, and electroencephalographic (EEG) measures, were examined as predictors of responder status.

Significant post-training improvement in cognitive composite score (Cohen’s d = .41) was observed across the sample, with n = 21 and n = 16 classified as responders and non-responders, respectively. Near learning was evidenced by significant improvement on each training exercise with practice; however, learning did not directly predict responder status. Group-wise comparison of responders and non-responders identified two factors favoring responders: higher EEG individual alpha frequency (IAF) and lower antipsychotic dosing. Tested in moderation analyses, IAF interacted with learning to predict improvement in cognitive outcome.

CR outcome in schizophrenia is not directly explained by learning during training and appears to depend on latent factors influencing far transfer of trained abilities. Further understanding of factors influencing transfer of learning is needed to optimize CR efficacy.