We investigated details of contrast TCD examination for detecting right-to-left shunts (RLS) in relation to respiratory maneuvers, especially the timing of a Valsalva maneuver (VM). We analyzed 14 patients with PFO proved by TEE. All underwent TCD (Multi-Dop X4-DWL) monitoring of the bilateral MCA. The following protocol with injections of 10mL of either agitated saline (PS)or agitated saline with one drop of Horizon® (ODM) as echo contrast; RM1: no VM, RM2: echo contrast injection during VM for 8seconds, RM3: VM for 5 seconds starting at the beginning of echo contrast injection, RM4: VM for 5 seconds 5 starting 5seconds after the beginning of echo contrast injection, RM5: VM for 5seconds starting 10seconds after the beginning of echo contrast injection, RM6: repetitive coughing between 2 and 13 seconds after the beginning of echo contrast injection. We counted contrast microembolic signals (cMES) for 60 seconds after the injection, and judged RLS to be positive if more than one cMES were detected. RM4 and RM5 with ODM, which demonstrated 12 cases to be RLS-positive, were the most sensitive maneuvers. Whichever test was used with PS, RLS was undetectable in no more than 8 cases. TCD with PS is an inadequate method for identifying RLS. The VM should be performed after the beginning of contrast injection.