Our study employed principles from Network Control Theory, an emerging field of engineering, to characterize the role of the left inferior frontal gyrus (LIFG) in controlling brain networks for word retrieval and selection under varying control demands.

Effective language production requires retrieving contextually appropriate words and selecting subsets of those words for spoken output. Broca’s area in the LIFG is implicated in controlled language production, but its specific role in controlled word retrieval and selection remains unclear. 

We examined two open-ended word generation tasks that manipulate retrieval and selection demands. Using fMRI data, we determined the degree to which each participant’s LIFG is suited for either (1) driving the network into states where modules are coupled or decoupled (i.e., boundary controllability) or (2) driving the network into difficult-to-reach states (i.e., modal controllability). We used continuous theta burst stimulation, a type of inhibitory transcranial magnetic stimulation (TMS), to observe how performance on selection and retrieval tasks is affected by TMS to the LIFG vs. the vertex (a control site). We examined relationships between controllability values in the LIFG and response times (RTs) before and after stimulation. 

In 28 healthy participants, individuals with higher LIFG boundary controllability exhibited practice-related interference (i.e., longer RTs post- vs. pre-vertex stimulation) for easy-to-retrieve items with many possible alternatives (i.e., high selection demands) and for hard-to-retrieve items with few alternatives (i.e., low selection demands). Active TMS over the LIFG eliminated this practice-related interference. There were no effects or interactions related to modal controllability. 

Thus, controlled retrieval and selection appear to rely on the integration and segregation between network states, and TMS can augment these processes in individuals with high LIFG boundary controllability. This suggests that we can optimize the application of neuromodulation for the remediation of post-stroke aphasia by characterizing aphasic patients’ language deficits and network controllability.