Over the last years, increasing evidence has fuelled the hypothesis that Autism Spectrum Disorder (ASD) is a condition of altered brain functional connectivity. The great majority of these empirical studies relies on functional magnetic resonance imaging (fMRI) which has a relatively poor temporal resolution. Only a handful of studies has examined networks emerging from dynamic coherence at the millisecond resolution and there are no investigations of coherence at the lowest frequencies in the power spectrum-which has recently been shown to reflect long-range cortico-cortical connections. Here we used electroencephalography (EEG) to assess dynamic brain connectivity in ASD focusing in the low-frequency (delta) range. We found that connectivity patterns were distinct in ASD and control populations and reflected a double dissociation: ASD subjects lacked long-range connections, with a most prominent deficit in fronto-occipital connections. Conversely, individuals with ASD showed increased short-range connections in lateral-frontal electrodes. This effect between categories showed a consistent parametric dependency: as ASD severity increased, short-range coherence was more pronounced and long-range coherence decreased. Theoretical arguments have been proposed arguing that distinct patterns of connectivity may result in networks with different efficiency in transmission of information. We show that the networks in ASD subjects have less Clustering coefficient, greater Characteristic Path Length than controls - indicating that the topology of the network departs from small-world behaviour - and greater modularity. Together these results show that delta-band coherence reveal qualitative and quantitative aspects associated with ASD pathology.

Anatomical and functional brain studies have converged to the hypothesis that Autism Spectrum Disorders (ASD) are associated with atypical connectivity. Using a modified resting-state paradigm to drive subjects' attention, we provide evidence of a very marked interaction between ASD brain functional connectivity and cognitive state. We show that functional connectivity changes in opposite ways in ASD and typicals as attention shifts from external world towards one's body generated information. Furthermore, ASD subject alter more markedly than typicals their connectivity across cognitive states. Using differences in brain connectivity across conditions, we classified ASD subjects at a performance around 80% while classification based on the connectivity patterns in any given cognitive state were close to chance. Connectivity between the Anterior Insula and dorsal-anterior Cingulate Cortex showed the highest classification accuracy and its strength increased with ASD severity. These results pave the path for diagnosis of mental pathologies based on functional brain networks obtained from a library of mental states.Running title: Brain networks in ASD are tuned for introspection.

Brain activity during wakefulness is characterized by rapid fluctuations in neuronal responses. Whether these fluctuations play any role in modulating the accuracy of behavioral responses is poorly understood. Here, we investigated whether and how trial changes in the population response impact sensory coding in monkey V1 and perceptual performance. Although the responses of individual neurons varied widely across trials, many cells tended to covary with the local population. When population activity was in a ‘low’ state, neurons had lower evoked responses and correlated variability, yet higher probability to predict perceptual accuracy. The impact of firing rate fluctuations on network and perceptual accuracy was strongest 200 ms before stimulus presentation, and it greatly diminished when the number of cells used to measure the state of the population was decreased. These findings indicate that enhanced perceptual discrimination occurs when population activity is in a ‘silent’ response mode in which neurons increase information extraction.

Information processing in the cerebral cortex depends not only on the nature of incoming stimuli, but also on the state of neuronal networks at the time of stimulation. That is, the same stimulus will be processed differently depending on the neuronal context in which it is received. A major factor that could influence neuronal context is the background, or ongoing neuronal activity before stimulation. In visual cortex, ongoing activity is known to play a critical role in the development of local circuits, yet whether it influences the coding of visual features in adult cortex is unclear. Here, we investigate whether and how the information encoded by individual neurons and populations in primary visual cortex (V1) depends on the ongoing activity before stimulus presentation. We report that when individual neurons are in a "low" prestimulus state, they have a higher capacity to discriminate stimulus features, such as orientation, despite their reduction in evoked responses. By measuring the distribution of prestimulus activity across a population of neurons, we found that network discrimination accuracy is improved in the low prestimulus state. Thus, the distribution of ongoing activity states across the network creates an "internal context" that dynamically filters incoming stimuli to modulate the accuracy of sensory coding. The modulation of stimulus coding by ongoing activity state is consistent with recurrent network models in which ongoing activity dynamically controls the balanced background excitation and inhibition to individual neurons.