ReAlnet: Achieving More Human Brain-Like Vision via Human Neural Representational Alignment