A Computational Model for Estimating NMDA Properties from Local Field Potential Spectra

Gabriele Mancini, Pablo Martínez-Cañada, Stefano Panzeri

Abstract: The balance between excitatory (E) and inhibitory (I) neural activity is crucial for brain function and is often disrupted in brain disorders. Recent model studies suggest that the slope of the frequency spectra of aggregate measures of neural activity, such as local field potentials (LFPs) or electroencephalograms (EEGs), can be used to predict the E/I ratio. However, these existing models overlook NMDA receptors, which are critical for cognitive functions like working memory and are implicated in disorders such as schizophrenia and Alzheimer’s. Here we use neural network simulations to show that the presence and strength of NMDA receptors have a major influence on the spectral slope of aggregate neural activity, particularly for lower frequencies, suggesting that spectral slopes of aggregate neural activity can be used to estimate NMDA receptor strength. We also find that estimates are better when using slopes computed from persistent activity after stimulus offset, a time when NMDA receptors affect most network dynamics. Moreover, when gamma oscillations are generated by the network, NMDA modulation affects the gamma peak spectral properties. The results suggest that the spectral slope of aggregate measures of neural activity such as LFPs or EEGs may be used as a biomarker of NMDA function or dysfunction.

External IDs:doi:10.1007/978-981-96-3294-7_30