Working memory (WM) maintenance of sensory information has been associated with enhanced cross-frequency coupling between the phase of low frequencies and the amplitude of high frequencies, particularly in medial temporal lobe (MTL) regions. It has been suggested that these WM maintenance processes are controlled by areas of the prefrontal cortex (PFC) via frontotemporal phase synchronisation in low frequency bands. Here, we investigated whether enhanced cognitive control during audiovisual WM as compared to visual WM alone is associated with increased low-frequency phase synchronisation between sensory areas maintaining WM content and areas from PFC. Using magnetoencephalography, we recorded neural oscillatory activity from healthy human participants engaged in an audiovisual delayed-match-to-sample task. We observed that regions from MTL, which showed enhanced theta-beta phase-amplitude coupling (PAC) during the WM delay window, exhibited stronger phase synchronisation within the theta-band (4-7 Hz) to areas from lateral PFC during audiovisual WM as compared to visual WM alone. Moreover, MTL areas also showed enhanced phase synchronisation to temporooccipital areas in the beta-band (20-32 Hz). Our results provide further evidence that a combination of long-range phase synchronisation and local PAC might constitute a mechanism for neuronal communication between distant brain regions and across frequencies during WM maintenance.