Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 21 (15), 1320-5

An in Vivo Assay of Synaptic Function Mediating Human Cognition


An in Vivo Assay of Synaptic Function Mediating Human Cognition

Rosalyn J Moran et al. Curr Biol.


The contribution of dopamine to working memory has been studied extensively [1-3]. Here, we exploited its well characterized effects [1-3] to validate a novel human in vivo assay of ongoing synaptic [4, 5] processing. We obtained magnetoencephalographic (MEG) measurements from subjects performing a working memory (WM) task during a within-subject, placebo-controlled, pharmacological (dopaminergic) challenge. By applying dynamic causal modeling (DCM), a Bayesian technique for neuronal system identification [6], to MEG signals from prefrontal cortex, we demonstrate that it is possible to infer synaptic signaling by specific ion channels in behaving humans. Dopamine-induced enhancement of WM performance was accompanied by significant changes in MEG signal power, and a DCM assay disclosed related changes in synaptic signaling. By estimating the contribution of ionotropic receptors (AMPA, NMDA, and GABA(A)) to the observed spectral response, we demonstrate changes in their function commensurate with the synaptic effects of dopamine. The validity of our model is reinforced by a striking quantitative effect on NMDA and AMPA receptor signaling that predicted behavioral improvement over subjects. Our results provide a proof-of-principle demonstration of a novel framework for inferring, noninvasively, neuromodulatory influences on ion channel signaling via specific ionotropic receptors, providing a window on the hidden synaptic events mediating discrete psychological processes in humans.


Figure 1
Figure 1
Experimental Task and Delay Period Activity (A) The experimental design comprised blocks of memory and no-memory trials. We ran six blocks of alternating memory and no-memory conditions, comprising 40 trials per block to yield a total of 240 trials per session. Drug and block order was counterbalanced across subjects. The first trial of each block was preceded by a cue indicating a memory or no-memory condition, and the background color was set to black for memory blocks and gray for no-memory blocks. Each trial consisted of a fixation cross (300 ± 50 ms) followed by the target stimulus of a colored square array. These arrays consisted of randomly colored squares (2.5° × 2.5°; red, blue, green, yellow, gray, cyan, and violet, with the number of squares corresponding to titrated load) in a randomized position. The target stimulus appeared for 300 ms and was followed by a delay period of 4000 ms, during which subjects had to retain the target memory array. This was followed by the onset of the probe image, after which subjects had 2000 ms to respond with a “match” or “no match” button press. Match and no-match trials occurred randomly, with equal probability. No-memory blocks contained the same stimuli, but subjects were instructed to simply press the “match” button on presentation of the second array. Accuracy was computed as the percentage of correct button press responses in memory blocks. Before drug administration on week 1, each subject was tested using 50 sample trials with varying memory load (two to six squares) to titrate accuracy levels. For each subject, the memory load with accuracy closest to 70% was used for the subsequent MEG experiment. (B) A pre-MEG test titrated individual memory load to achieve 70% accuracy. During MEG recordings, placebo-treated subjects performed with an accuracy close to titrated levels (70.60% ± 2.02% standard error of the mean [SEM]) and improved significantly on L-Dopa (74.04% ± 2.07%; p < 0.05, one-tailed paired t test). (C) We obtained scalp-time statistical parametric maps (SPMs) by testing for increases in sustained activity at particular bands during memory retention (see Supplemental Experimental Procedures). Sustained increases in delta and theta responses were found over prefrontal sites, whereas sustained alpha occurred primarily over occipitoparietal sites. (Beta activity, although significantly greater at the beginning of the delay period, did not show a sustained effect during maintenance, and no effect was expressed in the gamma band. All other bands showed sustained increases.) The images are maximum-intensity projection images showing t statistics for a significance level of p < 0.01 for clusters with more than ten pixels; color bars denote t values. In the top panel, SPMs of the t statistic are depicted for all sensors (left to right, corresponding to posterior to anterior) over time. The bottom panels depict these same (largest) statistical values over time with corresponding sensor locations. (D) Left: SPM of the interaction between memory and dopamine (displayed at p < 0.01 uncorrected), where the peak is observed in right superior frontal gyrus (SFG; peak x = 32, y = 4, z = 68; t = 2.79, p = 0.006 uncorrected within a mask of inferior, middle, and superior frontal gyri). The SPM is rendered on a canonical structural MRI scan, displayed on a horizontal section at z = 66. We also tested for the orthogonal main effect of memory and found significant increases in activity for memory compared to no-memory trials in bilateral prefrontal cortex, maximal over right dorsolateral prefrontal cortex (peak x = 22, y = 52, z = 0; t = 3.95, p = 0.001 uncorrected within a mask of inferior, middle, and superior frontal gyri; Figure S3). Middle: average spectral density for memory and no-memory trials, on L-Dopa (blue and red lines, respectively) and on placebo (green and cyan, respectively), in the right SFG, averaged across subjects (shaded regions report the SEM). Right: the interaction ([memory L-Dopa − no-memory L-Dopa] − [memory placebo − no-memory placebo]) plotted from the spectra (middle panel) averaged across subjects shows a negative theta response when retaining a memory on L-Dopa relative to placebo. This is effectively the difference in the differences among the four condition-specific responses. Bottom: macrocolumnar architecture used to model right SFG responses in the DCM analysis. The model comprises three interconnected cell layers, where spiny stellate cells occupy granular layer IV (population 1) whereas inhibitory interneurons (population 2) and pyramidal cells (population 3) occupy supra- and infragranular layers. For clarity, neurons in the infragranular layer are omitted. Extrinsic (e.g., thalamic) input enters the granular layer and signals propagate throughout the macrocolumn via intrinsic coupling parameters γto,from. The model's parameters are associated with particular ionotropic receptor types: AMPA and GABAA at all cell types, and NMDA at pyramidal cells and inhibitory interneurons. These synaptic parameters γ represent lumped coupling parameters that quantify the collective effect of a number of biophysical processes such as receptor binding and transmitter reuptake. The modeled SFG response is assumed to arise most prominently (80%) from the pyramidal cells' depolarization due to their dendritic organization, with a 20% contribution from the membrane potentials of the inhibitory interneurons and stellate cells.
Figure 2
Figure 2
Dynamic Causal Modeling Predictions and Parameter Estimates (A) Predicted and observed spectral responses from a single subject. Left: the responses of the memory dynamic causal modeling (DCM) showing predicted spectral responses with 90% Bayesian credible intervals for both L-Dopa and placebo conditions. These credible intervals include the observed spectrum at all frequencies. Right: observed and predicted spectral responses (again with 90% credible intervals) for no-memory trials. Again, the credible intervals of our model predictions include the observed spectra. (B) NMDA nonlinear function (Equation 2, Supplemental Experimental Procedures) illustrated for increasing values of parameter α. As α increases, the voltage-dependent magnesium switch becomes highly nonlinear. (C) Synaptic measures illustrating the difference in maximum a posteriori (MAP) B parameter estimates from the memory and no-memory DCMs. Significant differences were tested by one-tailed t test in hypothesized directions (p < 0.05) and were observed for two parameters, reflecting an increase in NMDA nonlinearity α and decreased exogenous (glutamatergic) input u under L-Dopa versus placebo, while subjects engaged in working memory relative to a control condition. (D) Left: using the difference in MAP B parameter estimates, individual differences in AMPA coupling γ1,3 show a significant negative correlation with behavioral improvement on L-Dopa (R = −0.46, p = 0.027, Pearson one-tailed linear correlation). Right: the interaction in NMDA nonlinearity α shows a significant positive correlation with behavioral improvement on L-Dopa (R = 0.55, p = 0.009, Pearson one-tailed linear correlation).

Similar articles

See all similar articles

Cited by 46 PubMed Central articles

See all "Cited by" articles


    1. Gao W.J., Goldman-Rakic P.S. Selective modulation of excitatory and inhibitory microcircuits by dopamine. Proc. Natl. Acad. Sci. USA. 2003;100:2836–2841. - PMC - PubMed
    1. Goldman-Rakic P.S. Regional and cellular fractionation of working memory. Proc. Natl. Acad. Sci. USA. 1996;93:13473–13480. - PMC - PubMed
    1. Robbins T.W. Chemical neuromodulation of frontal-executive functions in humans and other animals. Exp. Brain Res. 2000;133:130–138. - PubMed
    1. Kiebel S.J., David O., Friston K.J. Dynamic causal modelling of evoked responses in EEG/MEG with lead field parameterization. Neuroimage. 2006;30:1273–1284. - PubMed
    1. David O., Kiebel S.J., Harrison L.M., Mattout J., Kilner J.M., Friston K.J. Dynamic causal modeling of evoked responses in EEG and MEG. Neuroimage. 2006;30:1255–1272. - PubMed

Publication types