​Our daily activities depend on our ability to hold information in mind for immediate use. Until very recently, such working memory was thought to only rely on conscious brain processes. It was assumed that the contents of working memory as well as of the related operations (manipulating and transforming information) were conscious processes. Another very widely believed hypothesis was that information was maintained by persistent neural activity: the same neurons that initially encode a sensory stimulus remain active until that information is no longer needed. In other words, as long as you have to remember something, the cells in your brain that code for that information will be active. 

These predominant assumptions, however, were recently called into question. In 2011, it was claimed that invisible stimuli could be stored for several seconds - much longer than previously thought possible - by demonstrating that human participants responded better than chance when being asked to compare the orientation of an unseen stimulus shown seconds ago to one currently displayed on a computer screen . The existence of this long-lasting blindsight effect (i.e., above-chance objective performance in the absence of any conscious, subjective experience) led to the introduction of a novel concept of non-conscious working memory. 

In 2017, a collaboration led by Dr. Darinka Trübutschek and Professor Stanislas Dehaene at NeuroSpin then confirmed, for the first time, via MEG^[1] experiments the genuine nature of this long-lasting blindsight effect: contrary to century-old beliefs,   information could indeed be stored in working memory without subjects being conscious of it and, importantly, without the need for persistent brain activity. The non-conscious maintenance process rather relies on “activity-silent” mechanisms that are hidden to standard recording techniques. 

Until now, research had thus provided substantial evidence in favor of the non-conscious maintenance of information. But for there to exist a genuine non-conscious working memory system, information also needs to be  manipulated non-consciously. In their current study, the same NeuroSpin team, this time in collaboration with the Institut de Mathématiques de Jussieu, wanted to close this important gap and test whether such a non-conscious manipulation would indeed be possible. They therefore adapted their previous experiment by slightly modifying the general experimental setup to address this question. 

A barely visible square-shaped target was briefly flashed in one of 24 positions on a computer screen. Human volunteers had to locate the square and indicate whether they had seen it or not. They had to guess the location of the target whenever they had not seen it. In the new set-up, volunteers were asked to either just maintain or rotate the location of the subjectively seen or unseen target. While the volunteers performed this task, their brain activity was again continuously being monitored using MEG. 

Darinka Trübutschek explains “our observations may help to reconcile the current debate on the brain mechanisms of working memory by highlighting the complementarity of active and silent processes in working memory.” First, after a long delay, even when subjects had not seen the target at all, they still performed much better than predicted by chance when being asked to rotate it. But this long-lasting blindsight effect did not appear to result from a genuine non-conscious rotation: at the time subjects were asked to rotate the location of the target, neural signatures of consciousness emerged even on the unseen trials. Around the same time, the researchers also observed a transition from an activity-silent brain representation of the memorized unseen information to one coded by active neural firing. They thus observed that, while the short-term storage of information requires neither conscious nor persistent neural activity, manipulation is associated with both. As such, the existence of a genuine non-conscious working memory appears to be out of question. Instead, the authors propose an “activity-silent short-term memory” as an alternative for the phenomenon of long-lasting blindsight.