Abstract Mice – with their short gestation time, lifespan, large litter size, and above all, genetic tractability –are powerful experimental tools with which to explore mammalian brain development. One striking example is the reversible disruption of autism risk genes, like Mecp2 or Shank3. Switching on these genes after the emergence of fullblown phenotypes rescues several motor- related symptoms, including inertia, abnormal gait, weight, irregular breathing, repetitive grooming or hind-limb clasping. Instead, other neurological aspects are uncorrected in adulthood, such as anxiety or motor coordination. Of central relevance to human patients, it remains unknown to what extent higher cognition is impaired in Shank3 or Mecp2 mutant mice and whether recovery would also be limited to a critical period. The Hensch project will directly address sensitive periods and reversibility of dorsomedial prefrontal cortical (dmPFC) functions using novel behavioral tests of attention, cognitive flexibility and acoustic preference along with associated physiological measures from relevant areas. Pioneering work in the first phase of the Conte Center established the pivotal role of particular inhibitory neurons underlying critical period timing in mouse sensory systems. Parvalbumin (PV+) GABA circuit maturation dictates both the onset and closure of these windows of circuit refinement. Manipulations of psychiatric risk factors, such as circadian Clock gene disruption or redox dysregulation, can delay or extend developmental trajectories by upsetting the vulnerable PV+ component of local circuit excitatory-inhibitory balance. Recently, Hensch and Feng extended this principle to higher-order multi-sensory integration (MSI, commonly impaired in patients with autism) in the insular cortex. Shared MSI impairments in mice lacking either Shank3 (weak PV+) or Mecp2 (excessive PV+) suggest an optimal range of PV+ network function enables proper pruning of connections in the insula. Here, we will examine circuit physiology and anatomy before/after restoration of Shank3 or Mecp2 in mice, ultimately by full 3D EM circuit reconstruction with Lichtman also in marmosets carrying the same Shank3 deletion (from Feng). Further, our touchscreen two- choice visual attention assay, a multiple-choice foraging task to assess flexible rule learning, and preference for acoustic stimuli (music, ultrasonic calls) experienced early in life will probe dmPFC function in mutant and rescued mice. Electrophysiological recording and two-photon Calcium / Chloride imaging from the dmPFC in vivo will focus on PV+ networks in these areas across development, starting with comparison to Arlotta’s human organoids. Based on these many insights from mice, the impact of silencing/activating PV+ circuits in corresponding frontal cortical regions of PV-Cre marmosets (by Feng) using focal injections of viral DREADD constructs can be tested on analogous primate tasks of attention, cognitive flexibility and preference behavior at MIT. Our collective work will determine whether biological determinants of critical periods in sensory systems play a similar role in cognition, the corresponding circuit changes which are corrected when Shank3/Mecp2 symptoms are reversed and how much the mouse and primate brain differ.