BACKGROUND: Working memory (WM) is a key element of cognition that depends on accurate internal representation of stimuli and often impaired in psychiatric disorders, including schizophrenia. Prior human neuropharmacological studies with ketamine suggest that glutamate N-methyl-D-aspartate receptors (NMDARs) are critical for spatial WM (sWM), with NMDAR antagonism attenuating task-evoked neural activation and reducing WM accuracy. Animal studies, at the cellular level, and cortical microcircuit models hypothesize that NMDAR antagonism impairs sWM by decreasing the specificity of neuronal responses to spatial locations through broadened spatial tuning. Testing the relevance of this mechanism in humans remains a vital gap in our understanding of sWM and is critical for designing better treatments for WM impairments. METHODS: To overcome the challenges of measuring neuronal responses invasively in humans, we combined pharmacological manipulation and computational fMRI to show that ketamine broadens neural spatial tuning in healthy adults (N=40). Specifically, we modeled population neuron-level representations, rather than traditionally examined changes in blood oxygen level dependent (BOLD) activation. RESULTS: Broader neural tuning across visual perception and WM networks was linked to poorer sWM performance under ketamine, implicating tuning changes in sWM deficits. Furthermore, ketamine-induced changes in tuning, compared to the overall magnitude of BOLD activation, were more consistent across individuals and brain regions, suggesting a collective effect of ketamine on neural tuning across spatially-selective regions. CONCLUSIONS: These findings provide the first empirical evidence linking altered neuronal tuning to NMDAR antagonism and sWM impairments in humans, advancing a mechanistic understanding of glutamatergic mechanisms underlying sWM and related deficits in neuropsychiatric disorders.
