The functional brain networks that underlie visual working memory in the first two years of life

Abstract

Visual working memory (VWM) is a central cognitive system used to compare views of the world and detect changes in the local environment. This system undergoes dramatic development in the first two years; however, we know relatively little about the functional organization of VWM at the level of the brain. Here, we used image-based functional near-infrared spectroscopy (fNIRS) to test four hypotheses about the spatial organization of the VWM network in early development. Four-month-olds, 1-year-olds, and 2-year-olds completed a VWM task while we recorded neural activity from 19 cortical regions-of-interest identified from a meta-analysis of the adult fMRI literature on VWM. Results showed significant task-specific functional activation near 6 of 19 ROIs, revealing spatial consistency in the brain regions activated in our study and brain regions identified to be part of the VWM network in adult fMRI studies. Working memory related activation was centered on bilateral anterior intraparietal sulcus (aIPS), left temporoparietal junction (TPJ), and left ventral occipital complex (VOC), while visual exploratory measures were associated with activation in right dorsolateral prefrontal cortex, left TPJ, and bilateral IPS. Results show that a distributed brain network underlies functional changes in VWM in infancy, revealing new insights into the neural mechanisms that support infants' improved ability to remember visual information and to detect changes in an on-going visual stream.

Fig. 1

Experimental details and behavioral results. (a) Probe geometry laid over the sensitivity profile from an age-matched anatomical template. The figure depicts the regions of the brain we recorded from. Sources are marked with red circles; detectors are marked with blue circles. Channels are shown in green. Figure was created using AtlasviewerGUI (HOMER2, Massachusetts General Hospital/Harvard Medical School, MA, U.S.A.). (b) Schematic of a trial of the modified preferential looking task. The stimuli consisted of two side-by-side flickering displays composed of an array of colored squares, one side contained the change display and the other contained the no-change display. Each display contained two, four, or six colored squares. The squares simultaneously appeared for 500 ​ms and disappeared for 250 ​ms during the 10s trials. For the no-change display, the colors of the squares remained constant throughout the length of the trial. For the change display, one of the squares changed color after each delay. (c) Shift rate across set size. (d) Total looking time across ages. (e) Time course model fit to looking data from the task, indicating proportion of looks to the change side (change preference; CP) over time from trial onset. Points and point-ranges indicate means and standard errors of the data; lines indicate model fit. The grey dotted line indicates chance looking at a proportion of 0.5.

Fig. 2

Fit of the mixed ACF model to the empirical ACF in our fNIRS data. Green line depicts the canonical Gaussian ACF assumption, while black line shows the empirically estimated ACF values generated by the program 3dFWHMx. The red line shows the estimated mixed model after fitting parameters described in Cox et al. (2017).

Fig. 3

fNIRS ANOVA and linear contrast results. The line plots on the top panels show how the VWM network changed across ages in early development. Red lines/dots show HbO, blue lines/dots show HbR, shading depicts standard error. Panels show patterns of functional brain activity as a function of age in the left Ventral Occipital Cortex (VOC, A), the right Dorsolateral Prefrontal Cortex (DLPFC, B), and the left anterior Intraparietal Sulcus (aIPS, C). Brain images show significant clusters from the fNIRS ANOVA after familywise correction. Row D shows Hb and Age x Hb ANOVA results: pink ​= ​chromophore (Hb) effects, fuschia ​= ​Age x Hb effects, and brown ​= ​overlap between Hb and AgexHb effects. Row E shows Age x Hb general linear tests: mustard ​= ​4mo ​> ​1yo, and light green ​= ​1yo ​> ​4mo. ROIs from the adult fMRI literature are shown as teal circles.

Fig. 4

SS-related effects from the ANOVA and linear contrasts. The line plots in panel A shows patters of brain activity in right anterior Intraparietal Sulcus (aIPS) as a function of memory load (set size). Red lines/dots show HbO, blue lines/dots show HbR, shading depicts standard error. Panel B shows the Age x SS x Hb effect from the ANOVA: dark green ​= ​Age x SS x Hb effect; purple shows the significant cluster from the SS linear contrasts with SS med ​> ​SS high. ROIs from the adult fMRI literature are shown as teal circles.

Fig. 5

Relationships between change preference scores and functional brain activity. Panel A shows clusters in left VOC and left TPJ whose activity was significantly predicted by change preference scores. The line plots in the bottom panels show results from models predicting neural activity with behavior. Panel B shows the CP∗SS∗Hb interaction from l-TPJ (see Table 3), while panel C shows the same effect from l-VOC. Panel D shows the significant CP∗Age interaction in l-TPJ, while panel E shows the CP∗Age∗SS interaction in l-VOC. Colors are indicated by the legends. Lines and dots follow the same color scheme. In all line plots, shading depicts standard error.

Fig. 6

Relationships between brain activity and total looking time. Panel A shows clusters in left TPJ and right DLPFC whose activity was significantly predicted by total looking time. Panel B shows the TL∗Age interaction from l-TPJ (see Table 3) plotted for each chromophore separately for consistency with panel C. Panel C shows the TL∗Age effects from r-DLPFC, plotted separately for each chromophore to highlight the TL∗Hb effect in this region. Colors are indicated by the legend. Shading depicts standard error.

Fig. 7

Relationships between brain activity and shift rate. Panel A shows l-aIPS and r-aIPS clusters showing a significant relationship to shift rate over ages. Panels B (l-aIPS) and C (r-aIPS) show significant Shift Rate ​× ​Age interaction in linear models predicting brain activity from behavioral measures (see Table 3). Colors are indicated by the legend.