The neurofibrillary tangles associated with Alzheimer's disease first appear and attain their highest density in the entorhinal cortex. The identification of molecular pathways involved in Alzheimer's disease does not yet explain this selective sensitivity of the entorhinal cortex. My work focuses on the physiological properties of entorhinal cortex, some of which may be relevant to Alzheimer's disease. I will review modeling on two different levels of function: 1. Models of dynamical mechanisms for generation of single neuron physiological properties, and 2. Models of network dynamics that show a potential mechanism for the initiation and spread of Alzheimer's disease pathology and suggest pharmacological approaches to treatment using NMDA antagonists and muscarinic M4 agonists.

Single neuron recordings reveal grid cells in medial entorhinal cortex, that fire when a rat visits an array of locations in the environment (Moser and Moser, 2008). The spacing and size of firing fields is larger in grid cells recorded in more ventral anatomical locations (Sargolini et al., 2006). Models of grid cells using interference of oscillations predicted that this difference in spacing could arise from differences in the intrinsic oscillation frequency of entorhinal neurons (Burgess, Barry and O'Keefe, 2007). Whole cell patch data from my laboratory shows that neurons have higher frequencies of resonance and membrane potential oscillations in dorsal compared to ventral entorhinal cortex (Giocomo et al., 2007; Giocomo and Hasselmo, 2008), supporting the model. We further tested the role of oscillations by combining the recording of grid cells with inactivation of the medial septum by infusions of muscimol (Brandon et al. 2011). These infusions block theta rhythm oscillations in the entorhinal cortex and are accompanied by a loss of spatial periodicity of grid cell firing, while sparing head direction selectivity of entorhinal neurons. This supports an important role of theta rhythm oscillations in generating the spatially periodic firing of gird cells. Cholinergic modulation reduces resonance frequency of single neurons (Heys et al., 2010), providing a potential mechanism for changes of grid cell firing fields in novel environments. Another variant of the model (Hasselmo, 2008) uses the rhythmic persistent spiking induced in entorhinal cortex by muscarinic acetylcholine receptors (Fransen et al., 2006; Tahvildari et al., 2007). Newer versions of the model utilize network interactions between spiking neurons (Zilli and Hasselmo, 2010) or combine network attractor dynamics with oscillations (Hasselmo and Brandon, 2012).

In older research, I addressed network level models that show an interesting breakdown of function relevant to Alzheimer's disease (Hasselmo, 1994; 1997). In these models, interference between overlapping memories causes runaway synaptic modification of excitatory synapses in the entorhinal cortex and the hippocampus. This model of malignant synaptic growth provides a potential mechanism for the selective distribution of molecular pathology in terms of excessive demands placed on the remodeling of synaptic connections and on axonal transport for redistribution of synaptic resources. This network-level functional breakdown can spread between regions without requiring the transfer of molecular pathology. I will review this phenomenon in models and its potential contribution to molecular, anatomical and behavioral properties of Alzheimer's disease. This model shows how the blockade of NMDA receptors by the drug memantine could slow the spread of the pathology, and suggests that selective M4 receptor agonists could slow progression of pathology via selective presynaptic inhibition of glutamatergic transmission.