Laminar pattern of adolescent development changes in working memory neuronal activity

Abstract

Adolescent development is characterized by an improvement in cognitive abilities, such as working memory. Neurophysiological recordings in a nonhuman primate model of adolescence have revealed changes in neural activity that mirror improvement in behavior, including higher firing rate during the delay intervals of working memory tasks. The laminar distribution of these changes is unknown. By some accounts, persistent activity is more pronounced in superficial layers, so we sought to determine whether changes are most pronounced there. We therefore analyzed neurophysiological recordings from the young and adult stage of male monkeys, at different cortical depths. Superficial layers exhibited an increased baseline firing rate in the adult stage. Unexpectedly, we also detected substantial increases in delay period activity in the middle layers after adolescence, which was confirmed even after excluding penetrations near sulci. Finally, improved discriminability around the saccade period was most evident in the deeper layers. These results reveal the laminar pattern of neural activity maturation that is associated with cognitive improvement.NEW & NOTEWORTHY Structural brain changes are evident during adolescent development particularly in the cortical thickness of the prefrontal cortex, at a time when working memory ability increases markedly. The depth distribution of neurophysiological changes during adolescence is not known. Here, we show that neurophysiological changes are not confined to superficial layers, which have most often been implicated in the maintenance of working memory. Contrary to expectations, substantial changes were evident in intermediate layers of the prefrontal cortex.

Keywords: adolescence; cerebral cortex; layer; prefrontal cortex; working memory.

Graphical abstract

Figure 1.

A: sequence of events in the oculomotor delayed response (ODR) task. B: schematic diagram of the monkey brain with approximate location of neuronal recordings in areas 8 and 46 of the dorsolateral prefrontal cortex (dlPFC) highlighted. AS, arcuate sulcus; PS, principal sulcus.

Figure 2.

Recording depths. Depths of neurons with significant responses during the task relative to the top of the cortex are shown for the young (A) and adult (B) stages. Each horizontal line represents the depth of one neuron. Histograms summarize the depths of neurons in four monkeys in the young and adult stage. Blue horizontal lines indicate the 0, 800, and 1,200 μm depths that defined the boundaries for the superficial, middle, and deep groups.

Figure 3.

Firing rate in different depth groups and developmental stages. A: average, population peristimulus time histogram for neurons that responded to the visual stimulus and were recorded during the oculomotor delayed response (ODR) task from the superficial layer at adult and young stages. Responses are shown for a stimulus in the neuron’s receptive field and aligned to the stimulus onset of each trial. Vertical lines represent stimulus onset and fixation offset (n = 191 for the young, n = 212 for the adult stage). B: as in A, for the middle layer (n = 71 for the young, n = 54 for the adult). C: as in A, for the deep layer (n = 47 for the young, n = 56 for the adult). D–F: as in A–C, for trials with the best delay activity of each neuron. G–I: as in D–F, after excluding sulcal bank units (n = 126, 46, 25 for the young, n = 176, 43, 49 for the adult, in G, H, I, respectively).

Figure 4.

Receiver operating characteristic (ROC) analysis in each layer. A: mean area under the ROC curve in successive 100-ms windows is plotted as a function of time during the oculomotor delayed response (ODR) task, for superficial layer neurons at adult and young stages (n = 191 for the young, n = 212 for the adult stage). B and C: as in A, for the middle and deep layer, respectively (n = 71 for the young, n = 54 for the adult in middle; and n = 47 for the young, n = 56 for the adult in deep layers). D: percentages of neurons at adult (top) and young (bottom) stages reaching different levels of ROC values at each time point of the ODR task. E and F: as in D, for the middle layer and deep layer, respectively.

Figure 5.

Neuronal tuning in each layer. A: average activity (and means ± SE) during the cue period of the oculomotor delayed response (ODR) task in neurons recorded from the superficial layer at adult and young stages. Locations have been rotated, so that the best location of each neuron is represented in location 5. Location 9 is the same as location 1 (n = 191 for the young, n = 212 for the adult stage). Solid lines represent the best Gaussian fit of the population average. B and C: as in A, for the middle and deep layer, respectively (n = 71 for the young, n = 54 for the adult in middle; and n = 47 for the young, n = 56 for the adult in deep layers). D: average activity (and means ± SE) during the delay period of the ODR task from the superficial layer at adult and young stages. E and F: as in D, for the middle and deep layer, respectively.