The influence of sample and treatment characteristics on the efficacy of repetitive transcranial magnetic stimulation in depression: a meta-analysis and meta-regression