Basin structures in the northern Los Angeles area have been identified from a nodal seismic array along 10 lines across three basins. The dense array of 758 geophones spaced at 250-300 m apart along the lines and recorded seismic response for 30-35 days. For basin structure investigation, the teleseismic receiver functions technique was used. The primary basin concerted phases were identified from the receiver functions. A shear wave velocity model produced in a separate study using the same dataset was incorporated to convert the basin time arrivals into depth. The deepest part of the San Bernardino basin was identified near Loma Linda fault at 2.8 km. San Gabriel basin exhibit larger basin depths and Ps arrival times of all with a maximum depth of 4 km. The high lateral resolution from the dense array helped revealing more continuous structures and reducing uncertainties in the RFs. We discovered a more complex basin structure than previously identified. Our findings show the basins' core areas are not the deepest. Significant changes in basin depth were observed near some faults i.e. Rialto-Colton fault, Fontana fault, Red Hill fault and Raymond fault. Plain Language Summary Sedimentary basins can contribute to the damage caused by any earthquake. The seismic waves from an earthquake can get trapped and amplified within the basin which may result in a stronger ground shaking with longer duration. The three basins of our study area, San Bernardino, Chino and San Gabriel, are located in a seismically active region. The BASIN (Basin Amplification Seismic INvestigation) project aims to image the subsurface so that the seismic wave response within the basins for any earthquake can be assessed. As part of this project, we identified sedimentary basin shape and depth underneath this region. The P wave from an earthquake generates P to S conversions at boundaries. We used receiver functions to determine basin boundaries from those conversions. Arrival times of the P to S conversions were converted to depth using a velocity model.