The influence of position in the sleep wake cycle on the functional brain processes involved in memory consolidation during acquisition of an artificial language (BROCANTO)

Abstract

The neurological basis of learning an artificial language ‘BROCANTO’, and how this is influenced by sleep wake cycle position, was investigated using fMRI. BROCANTO requires both declarative (vocabulary) and procedural (grammar) processes. Behavioural results show equivalent performance can be reached with different declarative and procedural trajectories, with peak learning at 3pm and 24 hours consolidation. Right IFG, and left BA6 are key regions responding independently of performance, at this time. Left BA6 and right thalamus respond to the interaction between consolidation and morning/evening training. During rest, the connection between the hippocampus and IFG that is stronger in the morning and negative post task, especially on the right, indicating a BROCANTO learning related disconnection. IFG connectivity is particularly responsive to the experimental conditions during rest. In a PPI task analysis, the hippocampus input to the thalamus strengthens at the second session in the morning, and its connections to right IFG and BA6 decline, when its input to IFG is negative. Connections between IFG and BA6 and thalamus and BA6 are key. The connection between left BA6 and left IFG, responds to the interaction between consolidation and sleep/wake cycle, during task and rest. Together, the results suggest a hub role for BA6, modulated by the thalamus. This could represent an essential core BROCANTO processing system, mediating between declarative and procedural learning to be investigated further. The combined results do indicate that consolidation facilitates a shift towards procedural processing, and that its functional neural pathway is influenced by the sleep wake cycle. Analysis with further temporal information, and to identify strength and direction of connections would be needed to further clarify the mechanism. These findings can inform network models of language such as the ‘dorsal and ventral’ stream, with a view to developing optimal training schedules to benefit flexible study and work patterns.