Neuronal activity in human lateral temporal cortex related to short-term verbal memory, naming and reading.

Extracellular microelectrode recordings were obtained from lateral temporal cortex that was subsequently resected in patients undergoing craniotomies under local anaesthesia for treatment of medically intractable epilepsy. During these recordings patients performed visually presented measures of overt and silent naming and word reading, short-term verbal memory and a control task requiring matching of angles. These measures were designed so that the same visual stimuli elicited language, short-term memory or spatial responses. Statistically significant changes within and between these various measures were identified. Technically satisfactory recordings were obtained from 17 populations reflecting activity predominantly from 1 neuron, in 13 patients. Two populations demonstrated no significant changes in any measured functions. Only 1 population showed changes suggesting a relation to visual perception. Four populations in or adjacent to the superior temporal gyrus altered activity with overt speech. Four other populations in the anterior temporal lobe altered activity during silent, but not overt speech. Some relation to language or memory was established for 13 of the 17 populations: 1 altered activity during reading alone, 6 during memory alone, and 6 to both. Most of the recording sites showing these language and memory changes were not essential for those functions based on surface electrical stimulation mapping. Thus the area of temporal lobe that participates in language and memory, as indicated by changes in neuronal activity, is substantially larger than the areas essential for those functions as determined by stimulation mapping. Within that participatory area, changes related to language were most often an increase in activity sustained throughout the task, a pattern suggestive of mechanisms of selective attention. Changes related to memory most often included a sustained increase in activity at the time of entry of information into memory, and again at retrieval, with decreased activity during the time the memory was stored. A few neuronal populations demonstrated relative inhibition of activity during the memory task, compared with control measures.