process are formed in the auditory (temporal) cortex of the left hemisphere for most individuals. With further exposure, both the simple and complex circuits (corresponding to simple sounds and sequences of sounds) are activated at virtually the same time and more easily.

As connections are formed among adjacent neurons to form circuits, connections also begin to form with neurons in other regions of the brain that are associated with visual, tactile, and even olfactory information related to the sound of the word. These connections give the sound of the word meaning. Some of the brain sites for these other neurons are far from the neural circuits that correspond to the component sounds of the words; they include sites in other areas of the left hemisphere and even sites in the right hemisphere. The whole complex of interconnected neurons that are activated by the word is called a neural network.

The flow of neural activity is not unidirectional, from simple to complex; it also goes from complex to simple. For example, higher order neural circuits that are activated by contextual information associated with the word doggie can prime the lower order circuit associated with the sound doggie with the result that the word doggie can be retrieved with little direct input. Complex circuits can be activated at the same time as simple circuits, because the brain is receiving input from multiple external sourcesauditory, visual, spatial, motor. At the same time that the auditory circuit for the word doggie is activated, the visual circuit associated with the sight of a dog is also activated. Simultaneous activation of circuits in different areas of the brain is called parallel processing.

In early stages of learning, neural circuits are activated piecemeal, incompletely, and weakly. It is like getting a glimpse of a partially exposed and very blurry photo. With more experience, practice, and exposure, the picture becomes clearer and more detailed. As exposure is repeated, less input is needed to activate the entire network. With time, activation and recognition are relatively automatic, and the learner can direct her attention to other parts of the task. This also explains why learning takes time. Time is needed to establish new neural networks and connections between networks. This suggests that the neural mechanism for learning is essentially the same as the products of learninglearning is a process that establishes new connections among networks and the new skills or knowledge that are learned are neural circuits and networks.

What are the implications of these findings for teaching? First, effective teaching should include a focus on both parts and wholes. Instructional approaches that advocate teaching parts and not wholes or wholes and not parts are misguided, because the brain naturally links local neural activity to circuits that are related to different experiential domains. For example, in initial reading instruction, teaching phonics independently of the meaning of the words and their meaningful use is likely to be less effective than teaching both in parallel. Relating the mechanics of spelling to students’ meaningful use of written language to express themselves during diary writing, for example, provides important motivational incentives for learning to read and write. Second, and related to the preceding point, teaching (and learning) can proceed from the bottom up (simple to complex) and from the top down (complex to simple). Arguments