Neurons in primary visual cortex exhibit well documented centre-surround receptive field organization, whereby the centre is dominated by excitatory influences and the surround is generally dominated by inhibitory influences. These effects have largely been established by measuring the output of neurons, i.e. their spiking activity. How excitation and inhibition are reflected in the local field potential (LFP) is little understood. As this can bear on the interpretation of human fMRI BOLD data and on our understanding of the mechanisms of local field potential oscillations, we measured spatial integration and centre-surround properties in single- and multiunit recordings of V1 in the awake fixating macaque monkey, and compared these to spectral power in different frequency bands of simultaneously recorded LFPs. We quantified centre-surround organization by determining the size of the summation and suppression area in spiking activity as well as in different frequency bands of the LFP, with the main focus on the gamma band. Gratings extending beyond the summation area usually inhibited spiking activity while the LFP gamma-band activity increased monotonically for all grating sizes. This increase was maximal for stimuli infringing upon the near classical receptive field surround, where suppression started to dominate spiking activity. Thus, suppressive influences in primary cortex can be inferred from spiking activity, but they also seem to affect specific features of gamma-band LFP activity.

The functional magnetic resonance imaging (fMRI) blood oxygenation level-dependent (BOLD) signal is regularly used to assign neuronal activity to cognitive function. Recent analyses have shown that the local field potential (LFP) gamma power is a better predictor of the fMRI BOLD signal than spiking activity. However, LFP gamma power and spiking activity are usually correlated, clouding the analysis of the neural basis of the BOLD signal. We show that changes in LFP gamma power and spiking activity in the primary visual cortex (V1) of the awake primate can be dissociated by using grating and plaid pattern stimuli, which differentially engage surround suppression and cross-orientation inhibition/facilitation within and between cortical columns. Grating presentation yielded substantial V1 LFP gamma frequency oscillations and significant multi-unit activity. Plaid pattern presentation significantly reduced the LFP gamma power while increasing population multi-unit activity. The fMRI BOLD activity followed the LFP gamma power changes, not the multi-unit activity. Inference of neuronal activity from the fMRI BOLD signal thus requires detailed a priori knowledge of how different stimuli or tasks activate the cortical network.

Neural activity fluctuates endogenously on timescales varying across the neocortex. The variation in these intrinsic timescales relates to the functional specialization of cortical areas and their involvement in the temporal integration of information. Yet, it is unknown whether the timescales can adjust rapidly and selectively to the demands of a cognitive task. We measured intrinsic timescales of local spiking activity within columns of area V4 while monkeys performed spatial attention tasks. The ongoing spiking activity unfolded across at least two distinct timescales---fast and slow---and the slow timescale increased when monkeys attended to the receptive fields location. A recurrent network model shows that multiple timescales in local dynamics arise from spatial connectivity mimicking vertical and horizontal interactions in visual cortex and that slow timescales increase with the efficacy of recurrent interactions. Our results reveal that targeted neural populations integrate information over variable timescales following the demands of a cognitive task and propose an underlying network mechanism.