Disrupted serial dependence suggests deficits in synaptic potentiation in anti-NMDAR encephalitis and schizophrenia

Abstract

We report markedly reduced working memory-related serial dependence with preserved memory accuracy in anti-NMDAR encephalitis and schizophrenia. We argue that NMDAR-related changes in cortical excitation, while quickly destabilizing persistent neural activity, cannot fully account for a reduction of memory-dependent biases. Rather, our modeling results support a disruption of a memory mechanism operating on a longer timescale, such as short-term potentiation.The NMDA receptor (NMDAR) subserves memory mechanisms at several timescales, including sustained working memory delay activity 1,2 and different temporal components of synaptic potentiation 3,4 . In addition, hypofunction of NMDARs is linked to psychiatric disease, in particular schizophrenia 5 , and it possibly contributes to abnormal working memory function in patients with schizophrenia 6,7 . Indeed, reduced prefrontal NMDAR density characterizes this disease 8 . Yet, the specific neural alterations by which NMDAR hypofunction could lead to memory deficits in schizophrenia are still under debate 6,7 . Here, we studied working memory function in healthy controls, patients with schizophrenia, and patients recovering from anti-NMDAR encephalitis ( Methods , Supplementary Table 1). Anti-NMDAR encephalitis is characterized by an antibody-mediated reduction of NMDARs, accompanied by initial psychosis and long-lasting memory deficits 9,10 , resembling clinical features of schizophrenia 11 .Consequently, we expected working memory deficits in anti-NMDAR encephalitis to parallel those in schizophrenia. This correspondence allows linking alterations in working memory to the NMDAR in both patient groups.We assessed memory alterations in a visuospatial delayed-response task (Fig. 1a) on two coexisting temporal scales: single-trial working memory accuracy as a proxy of active memory maintenance during short delays, and serial dependence of responses on previously memorized stimuli 12,13 (serial biases, Fig. 1b) as a read-out of passive information maintenance across trials. Neural correlates of this task have been identified in monkey prefrontal cortex 14,15 , inspiring computational models that can capture key aspects of neural dynamics and behavior [15][16][17][18][19] . The biophysical detail of these models permits to investigate how NMDAR hypofunction at different synaptic sites affects circuit dynamics and working memory.Candidate mechanisms are a disturbed balance between cortical excitation and inhibition (E/I balance), as it is observed in schizophrenia and in studies using NMDAR antagonists (e.g. ketamine) 2,5,20,21 , and alterations in NMDAR-regulated short-term synaptic potentiation 3,4,22 . In the modeling section of this study (Figs. 2,3), we systematically tested the potential of these candidate mechanisms for explaining experimentally observed memory alterations in schizophrenia and anti-NMDAR encephalitis.First, we sought to identify alterations in single-trial working memory accuracy, as an indication of a possible dysfunction of activity-based memory...

Interplay between persistent activity and activity-silent dynamics in the prefrontal cortex underlies serial biases in working memory

“…Fig. 5 ; see 31 , 42 for an alternative mechanism). We finally tested the dependence of this behavioral effect on the strength of the nonspecific drive.…”

“… 11 , 17 – 19 ), with the cost of across-trial interference of memories 11 , 17 . Along these lines, a recently found reduction in serial biases in patients with schizophrenia 42 , anti-NMDAR encephalitis 42 or autism 28 may reflect a reduced interplay of memory supporting mechanisms. Second, we see an active role of PFC in generating serial biases, rather than suppressing them as proposed by the proactive interference literature 29 , 30 .…”

Task feedback suggests a post-perceptual component to serial dependence

“…There are multiple proposals for how serial dependence may arise from post-perceptual processes. The bump-reactivation hypothesis (Barbosa et al, 2020;Stein et al, 2020) proposes that serial biases originate in working memory through the reactivation of latent activity-silent patterns of neuronal activity related to the last representation from the previous trial, upon initiation of the current trial. As argued by Ceylan et al (2021), this model can account for biases with respect to previous task-relevant decisions, in our case, the previous reported 3D motion direction, because the decision is likely to be the last representation held in working memory when a trial ends.…”

A practical guide for studying human behavior in the lab

“…Depending on your question (e.g. studying population differences of serial dependence in visual processing (Stein et al, 2020) vs. studying the impact of visual serial dependence on emotion perception (Chen & Whitney, 2020)) it might be wise to use naturalistic stimuli, rather than synthetic ones.…”