Local cerebral glucose utilization after chronic electroconvulsive shock: implications for the mode of action of electroconvulsive therapy