This schematic illustrates how, within a 4 Hz cycle, the phase specificity and relative influence of dmPFC assembly and non-assembly neurons (AN, ON, respectively) could represent a functional signal for downstream structures implicated in the expression of freezing behaviour. a, In normal freezing conditions, the prevalence and restriction of AN activation in the ascending phase combined with the monotonic phase distribution of ONs provides downstream structures with a signal-to-noise ratio (SNR) over an entire 4 Hz cycle that allows freezing expression. b, During optogenetic PN inhibition this SNR is altered. Top: when stimulating in the ascending phase, the bulk of AN activity over a cycle is shut down. This dramatically reduces SNR and prevents freezing expression. Bottom: when stimulating in the descending phase, AN activity is largely preserved over a cycle while that of ONs is diminished to a large extent. This strongly increases SNR and exacerbates freezing expression. c, During optogenetic PN disinhibition the SNR is also altered. Top: when stimulating in the ascending phase, the AN activity over a cycle is promoted compared with non-ensemble activity. This increases SNR and enhances freezing expression. Bottom: when stimulating in the descending phase, both assembly and non-assembly neuron activities over a cycle are increased. This does not affect SNR and freezing remains similar to normal conditions. In this model the dmPFC can be seen as an emitter whose signal transmission conditions the SNR for the receiver and the further expression of fear behaviour. Our optogenetic dissection of phase specificity for dmPFC neuronal assemblies shows that SNR can be manipulated to either.