Local mAChR activation via top-down attentional signals is also important in our model for facilitating top-down attention in V1 and helps to both increase the firing rate and decrease noise correlations between these neurons (Herrero et al., 2008; Goard & Dan, 2009). Specifically, our model highlights how mAChR stimulation of excitatory neurons is important for attentional modulation AZD6244 mw while mAChR stimulation of inhibitory neurons is important for maintaining low levels of excitatory–excitatory correlations when

excitatory drive is increased. Contrary to recent experimental studies, which suggest a decrease in excitatory–excitatory correlations between neurons with BF stimulation and top-down attention, our model indicates that

attention and mAChR stimulation in V1 lead to a decrease in excitatory–inhibitory correlations, but cause no change in excitatory–excitatory correlations. Thus, because it is difficult to distinguish between excitatory and inhibitory neurons experimentally (Nowak et al., 2003; Vigneswaran et al., 2011), it is possible that experimenters are seeing excitatory–inhibitory rather than Selleckchem Copanlisib excitatory–excitatory decorrelations. This is a strong prediction of our model. We suggest inhibition may act as a mechanism for absorbing additional excitatory input that may result from increased excitatory drive from top-down attentional signals or Lepirudin activation of mAChRs on excitatory neurons in order to extinguish excess excitatory–excitatory correlations. A model was developed that contained two cortical columns, simulating two receptive fields, and was subject to both neuromodulation by the BF and top-down

attention (see Fig. 3). Input to the model was a movie of a natural scene as described below. Our goal was to see how neuromodulatory and top-down attention signals interacted and influenced between-trial and between-neuron correlations in the simulated cortical columns. Our experiment consisted of 60 trials, in which a 12-s natural scene video was input to the spiking neural network. We used this natural stimulus because it is similar to that used in Goard & Dan’s (2009) experiments and affords comparison of our model’s responses with their results. The video was obtained from the van Hateren movie database to the network (http://biology.ucsd.edu/labs/reinagel/pam/NaturalMovie.html). Experiments consisted of six blocks of ten trials (see Fig. 2A). In each block of ten trials, five were performed without BF stimulation, top-down attention and/or mAChR stimulation (control) followed by five trials with BF stimulation, top-down attention and/or mAChR stimulation (non-control). In between each trial and block, 1 and 4 s, respectively, of random, Poissonian spikes was injected into the network at a rate of 2 Hz to allow network activity to settle. The total simulation time of the experiment was 13.4 min.