Early‐SEPs’ amplitude reduction is reliable for poor‐outcome prediction after cardiac arrest?