Aim. In neuroscience, PET functional imaging and behavioural assays in rodents are complementary approaches, despite the fact that they are rarely associated simultaneously because general anaesthesia inherent to PET precludes behavioural studies. To address this methodological limit, we have developed a radiosensitive pixelated intracerebral probe, PIXSIC, that provides access to the combination of simultaneous observations of molecular and behavioural parameters on rodents. Material and Methods. PIXSIC proposes a novel strategy for in vivo recording of the local time-activity curves of PET radiopharmaceuticals. It relies on a sub-millimetre pixelated probe of Si (17 mm long hosting 10 pixels with dimension 200 µm x 500 µm) implanted into the brain region of interest by stereotaxic surgery. Positrons are detected by reverse-biased, high-resistivity silicon diodes. The system Aims at time-resolved high sensitivity measurements in a volume of a few mm3. The pixelated detection scheme adds "imaging" features as it allows recording of the time-activity curves in different brain regions along the probe position. PIXSIC has a compact and autonomous design based on a radiofrequency data exchange link that allows for full freedom in the animals motion and behavioural activity while limiting stress during acquisition. Results and Conclusion. The first biological validations were performed on anaesthetized rats implanted with two probes, one in the region of interest (hippocampus or striatum, according to the radiotracer) and the other one in a control region (cerebellum). We used [11C]-raclopride for dopamine D2 receptors and [18F]-MPPF for serotonin 5HT1A receptors. According to our previous studies with the Beta-Microprobe (J Nucl Med 2002, 43(2):227-33; Eur J Nucl Med 2002 29(9) 1237-47), the radioactive signals measured with the PIXSIC pixels are reproducible and well-correlated with the distributions of the targeted receptors. The simultaneous measurement of implanted rats in a small animal PET camera confirmed the similarity between PIXSIC and microPET time-activity curves. Moreover, the binding curves highlighted the possibility for PIXSIC to distinguish different tracer kinetics within the structure of interest (cortex/striatum or cortex/hippocampus) in accordance to the stereotaxic location of the pixels. In addition, PIXSIC allowed us to perform the first kinetic measurements of [11C]-raclopride and [18F]-MPPF on awake and freely moving rats. In conclusion, PIXSIC constitutes an unprecedented instrumental methodology for connecting PET molecular imaging and behavioral measurements with freely-moving rodents.