A sufficient and necessary NCC (or perhaps a set of NCCs) should be generic enough to cover all conscious contents and states, and should also be specific enough to cover only conscious contents and states. Notably, gamma synchronisation does not meet the second specificity requirement, as it can be associated with almost any perceptual and cognitive function that depends on the formation of temporary associations of distributed neuronal networks. Among numerous cases, increased gamma-band synchronisation is found to be associated with such tasks as perceptual learning (Gruber et al. 2002), self-paced movement (Pfurtscheller et al. 2003), mental rotation (Bhattacharya et al. 2001), viewing of unpleasant stimuli (Martini et al. 2012), deductive reasoning (Zhang et al. 2014), auditory attention control (Doesburg et al. 2012), face integration (Kottlow et al. 2012), and memory encoding and retrieval (Osipova et al. 2006). Gamma-band synchronisation thus seems be a generic process that contributes to complex cortical computations involved in most if not all of the higher cognitive functions (Fries 2009).
Singer (this collection) proposes that only global, widespread synchronisation of gamma oscillations is associated with consciousness, whereas local, spatially-restricted synchronisation is not necessary related to conscious awareness. This might refute studies reporting local synchrony; however, some of the above-mentioned studies found increased global gamma-synchrony when observing unpleasant stimuli (Martini et al. 2012) or carrying out a mental rotation task (Bhattacharya et al. 2001). It could be argued, though, that gamma synchronisation is present in these experiments as a correlate of task-related subjective experiences, such as awareness of memory retrieval. In fact, even though most of these studies did not even mention consciousness or awareness, their participants were not unconscious, and gamma synchronisation could have been associated with the task-dependent subjective experiences. Yet, this line of reasoning is challenged by the simple fact that participants remained conscious in all contrast conditions throughout the experiments. Why would consciousness-related gamma synchronisation increase in some, but not other conditions? This leads us to the question of what exactly gamma synchrony correlates with in studies that specifically manipulate awareness? Let us take a closer look at two key studies, also examined by Singer (this collection).
Melloni et al. (2007) presented pairs of words and asked participants to report on whether both words were the same. Visibility of the first word was manipulated by adjusting the luminance level of the forward and backward masks, which rendered the words visible only in some of the trials. Global gamma-phase synchronisation between the fronto-centro-parietal electrodes was observed within the 40–182ms time-window after the presentation of the first word only in visible trials, which coincides with the time when conscious perception of the words is expected to emerge. In the latter time-windows, visible words were marked by more localised gamma synchronisation, higher P300 amplitude, and higher amplitude of frontal theta oscillations than invisible words. These findings confirmed that gamma synchronisation is a correlate of visual-semantic awareness, and showed that other electrophysiological processes may also correlate with consciousness.
Doesburg et al. (2009) investigated the role of gamma-phase synchronisation in conscious awareness using a binocular rivalry paradigm, in which a different visual stimulus is presented to each eye. Instead of seeing both stimuli at the same time, people report perceiving only one of the stimuli that continues switching in time. An increase in the gamma-band synchronisation over the fronto-parietal regions was observed in the 600–540ms and 280–220ms time-windows before responses indicating a perceptual switch. Assuming that reaction time was about 250ms, the synchronisation increase coincided with a new percept reaching awareness. Interestingly, gamma synchronisation oscillated at the theta rhythm, suggesting a cross-frequency interaction.
In both of these experiments (Doesburg et al. 2009; Melloni et al. 2007), gamma synchrony peaked around the time when participants began experiencing a new content of consciousness, following which Singer (this collection) draws the well-justified conclusion that gamma synchronisation is associated with a transfer of the new contents into awareness. Given that increased synchronisation may reflect the neural and phenomenal binding required for the fundamental unity of consciousness to emerge, it seems to be an ideal candidate for the NCC. Yet the duration of increased synchronisation is relatively brief and seems to last a much shorter time than the awareness of stimuli. For instance, P300 distinguished visible and invisible words around 300ms post-stimulus, whereas gamma-band synchronisation became local during this time-window (Melloni et al. 2007). Such brevity of synchronisation suggests that it is involved only in the initial binding of the new contents of consciousness, while a further maintenance of these contents is supported by other neural mechanisms, in particular theta oscillations (Doesburg et al. 2009; Melloni et al. 2007; Singer this collection). Given that the global gamma synchronisation correlates with a spatially- and temporally-local change in the stream of consciousness, its association with an overall unity of consciousness is uncertain and, at least currently, it cannot be accounted as the only or even as the major NCC. If it were such, it would not cease as long as the participant were aware of a particular content of consciousness.
Furthermore, gamma synchrony does not seem to increase in response to each of the new contents of consciousness. In each trial, Melloni et al. (2007) presented a series of stimuli, including a fixation cross, a masking noise, a target word, and a blank screen. Each of these stimuli should have entered consciousness, and even when the target word was unreadable, participants should have perceived something, e.g., an unreadable word, incoherent letters, or a flashing mask. However, the global gamma synchronisation increased only in response to perceived visible words, suggesting that it is a correlate of the initial binding of a selected, expected, attended, coherent, task-relevant content of consciousness. As such, in addition to the lack of specificity, gamma synchrony does not seem to be generic enough to cover all the different contents of consciousness, even within the paradigmatic visual modality. Thus, it seems that the global gamma synchronisation is neither necessary not sufficient for consciousness to emerge, as subjective experiences may exist without gamma synchronisation, and even when synchronisation is involved in the generation of awareness, other neural processes are needed to maintain its presence.
As discussed in the previous section, the unity of consciousness emerges from the interaction of all experiences available at a time. Gamma synchronisation cannot account for the unity of a state of consciousness, simply because it is involved only in the generation of new task-dependent contents, and it does not seem to bind these contents within the broader stream of consciousness. Arguably, instead of focusing on selected stimuli, we may be able to detect the neural correlates of the unity of consciousness by contrasting states of consciousness with unconsciousness, since such a contrast would consider the whole stream of phenomenal contents, including their structural unity.