The recent publication by Wang and colleagues provides thought-provoking data about the role of NMDA receptors in “higher” cognitive functions such as working memory, and the potential role that NMDA receptor dysfunction might play in psychiatric illnesses such as schizophrenia. Briefly, Wang and colleagues show that both direct iontophoresis and systemic administration of NMDAR antagonists reduced delay-related firing of cells in deep layer III of primate dorsolateral prefrontal cortex (DLPFC). This effect occurred for both antagonism of NR2B and NR2A receptors, though the focus of Wang and colleagues' report was on NR2B receptors, given their localization to the postsynaptic density and the fact that the temporal dynamics of their kinetics make them good candidates for playing a specific role in the maintenance of persistent firing in DLPFC circuits, as explicated by computational modeling.
The results of Wang and colleagues were consistent with a computational model of the role of NR2B receptors in working memory and delay-related activity in DLPFC neurons (
The results of Wang and colleagues were consistent with a computational model of the role of NR2B receptors in working memory and delay-related activity in DLPFC neurons (Wang, 2001; Wang et al., 2008). One of the most important implications for these results in terms of understanding the potential role of NMDA dysfunction in schizophrenia is that they suggest that NMDA dysfunction can lead to a reduction in pyramidal cell firing during critical task conditions, such as when information needs to be actively maintained in working memory. This is in contrast to the more traditional focus of hypotheses about NMDA dysfunction in schizophrenia, which have highlighted a role for NMDAR blockade in decreasing inhibition and in the sculpting of DLPFC circuit activity, mediated via GABAergic inhibition on the firing of pyramidal cells. Of course, such disinhibitory effects may still be a critical part of the process by which NMDAR dysfunction could lead to cognitive or other impairments, but these data do suggest that the effects of NMDAR dysfunction extend beyond effects on GABA cells, and that working memory dysfunction associated with NMDAR antagonism may not be solely the result of disinhibition.
An intriguing feature of these data is the fact that while direct iontophoresis of antagonists for both NR2B and NR2A receptors reduced firing of cue and response cells as well as delay cells, systemic administration of ketamine reduced activity in delay-related cells, but also increased activity in response cells. The fact that this increase in response cell activity occurred only with systemic administration led Wang and colleagues to suggest that it may result from extra-PFC effects of systemic NMDAR blockade, or at least effects beyond the specific PFC column influenced by their iontophoresis method. They argue that these results might help explain why administration of systemic NMDAR antagonists in humans can manifest as increased BOLD response, given that response cells (located in layer V) are large and highly prevalent in PFC, and thus may strongly influence BOLD fMRI findings. However, this should only be true if the working memory paradigm does not allow the separation of activity associated with cue, delay and response periods (e.g., N-back tasks that may “average” across activity associated with all three task components). In contrast, there is increasing use in both the schizophrenia (Johnson et al., 2006; Driesen et al., 2008; Edwards et al., 2010) and the human pharmacology (Anticevic et al., 2012) literatures of paradigms—frequently referred to as delayed-match-to-sample tasks—that do allow this separation.
The recent work by Anticevic et al. that examined working memory related activity under ketamine using such a delayed-match-to-sample task showed clear evidence of decreased activity maintenance-related activity (e.g., delay activity) in DLPFC, which they interpreted in the framework of the disinhibition theory of the effect on NMDAR antagonism on local DLPFC cortical circuitry. However, these new data by Wang et al. suggest that these same findings might instead (or in addition) reflect direct effects of NMDAR blockade on pyramidal cell firing. Further, all of the studies using delayed-match-to-sample tasks in schizophrenia have also found reduced maintenance-related activity in DLPFC regions (Johnson et al., 2006; Driesen et al., 2008; Edwards et al., 2010; Anticevic et al., 2012). Such findings in schizophrenia would be consistent with a hypothesis of altered NR2B function that disrupts delay-related activity in DLPFC.
At the same time, the studies using delayed-match-to-sample paradigms have provided some results that are not fully consistent with the data presented by Wang et al. Specifically, several of the studies in schizophrenia and the study in healthy volunteers on ketamine did not find evidence of increased activity during the response period in DLPFC. Work from our lab did show evidence of increased response-related activity in schizophrenia during a variant of a working memory task that focused on the maintenance of context information, but we interpreted this increased response-related activity as reflecting the need to engage reactive control due to deficits in proactive control (Edwards et al., 2010). In addition, Anticevic et al. did not find any evidence of an influence of systemic ketamine on probe- or response-related activity in healthy humans. Honey and colleagues (Honey et al., 2004) did find an increase in brain activity in both PFC and parietal regions during working memory as a function of ketamine, but they used an analysis that focused on the cue and initial delay period in a working memory task. As such, one would have been predicted to show reduced activity—at least in DLPFC—based on the results of Wang and colleagues. However, ketamine only increased working memory-related brain activity in a manipulation versus maintenance contrast, and not in a high versus low load contrast. Thus, it is possible that the manipulation condition engages response processes even during the delay period.
Overall, the elegant data presented by Wang and colleagues are highly intriguing and offer an alternative way in which to understand and model the effects of abnormal NMDAR function on working memory and DLPFC function. At the same time, more work is needed to understand the degree to which findings in humans under NMDAR antagonism or in humans with illnesses such as schizophrenia correspond to the predictions made from the influence of NMDAR receptors on pyramidal cell firing versus interneurons. This work should take into account the potential differences between very local activity measured as spike rates in single cells versus fMRI BOLD data, which are an indirect measure of neural activity that aggregates across much larger spatial scales. In addition, this work will need to take into account potential effects of age, illness duration, and medication, all of which may change the dynamics and effects of potential NMDAR dysfunction or modulation.
Anticevic A, Gancsos M, Murray JD, Repovs G, Driesen NR, Ennis DJ, Niciu MJ, Morgan PT, Surti TS, Bloch MH, Ramani R, Smith MA, Wang XJ, Krystal JH, Corlett PR. NMDA receptor function in large-scale anticorrelated neural systems with implications for cognition and schizophrenia. Proc Natl Acad Sci U S A . 2012 Oct 9 ; 109(41):16720-5. Abstract
Anticevic A, Repovs G, Barch DM. Working memory encoding and maintenance deficits in schizophrenia: neural evidence for activation and deactivation abnormalities. Schizophr Bull . 2013 Jan ; 39(1):168-78
Driesen NR, Leung HC, Calhoun VD, Constable RT, Gueorguieva R, Hoffman R, Skudlarski P, Goldman-Rakic PS, Krystal JH. Impairment of working memory maintenance and response in schizophrenia: functional magnetic resonance imaging evidence. Biol Psychiatry . 2008 Dec 15 ; 64(12):1026-34. Abstract
Edwards BG, Barch DM, Braver TS. Improving prefrontal cortex function in schizophrenia through focused training of cognitive control. Front Hum Neurosci . 2010 ; 4():32. Abstract
Honey, R. A., G. D. Honey, et al., (2004). "Acute ketamine administration alters the brain responses to executive demands in a verbal working memory task: an FMRI study." Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology 29(6): 1203-1214.
Johnson MR, Morris NA, Astur RS, Calhoun VD, Mathalon DH, Kiehl KA, Pearlson GD. A functional magnetic resonance imaging study of working memory abnormalities in schizophrenia. Biol Psychiatry . 2006 Jul 1 ; 60(1):11-21. Abstract
Wang H, Stradtman GG, Wang XJ, Gao WJ. A specialized NMDA receptor function in layer 5 recurrent microcircuitry of the adult rat prefrontal cortex. Proc Natl Acad Sci U S A . 2008 Oct 28 ; 105(43):16791-6. Abstract
Wang XJ. Synaptic reverberation underlying mnemonic persistent activity. Trends Neurosci . 2001 Aug ; 24(8):455-63. Abstract