Assessing infants' cortical response to speech using near-infrared spectroscopy

Abstract

Sensitivity to spoken language is an integral part of infants' formative development, yet relatively little is known about the neural mechanisms that underlie the emerging ability to perceive and process speech. This is in large part because there are a limited number of non-invasive techniques available to measure brain functioning in human infants. Near-infrared spectroscopy (NIRS), an optical imaging technique that estimates changes in neuronal activity by measuring changes in total hemoglobin concentration and oxygenation, may be a viable procedure for assessing the relation between speech processing and brain function in human infants. While auditory processing data have been gathered from newborn and preterm infants using NIRS, such data have not been collected from older infants. Many behavioral measures used to establish linguistic sensitivity in this population are accompanied by visual stimuli; however, it is unclear how coupling of auditory and visual stimuli influences neural processing. Here we studied cortical activity in infants aged 6-9 months, as measured by NIRS, during exposure to linguistic stimuli paired with visual stimuli and compared this to the activity observed in the same regions during exposure to visual stimuli alone. Results dissociate infants' hemodynamic responses to multimodal and unimodal stimuli, demonstrating the utility of NIRS for studying perceptual development in infants. In particular, these findings support the use of NIRS to study the neurobiology of language development in older infants, a task that is difficult to accomplish without the use of attention-getting visual stimuli.

Figure 1

Infant fitted with the NIRS probe (a); infant oriented towards visual stimuli in the testing booth (b).

Figure 1

Infant fitted with the NIRS probe (a); infant oriented towards visual stimuli in the testing booth (b).

Figure 2

Probe design for two regions of interest (a); localization of the probe over T₃ (for left temporal measurements) and over O₁ and O₂ (for occipital measurements) was guided by International 10-20 EEG electrode position classification system (b).

Figure 2

Probe design for two regions of interest (a); localization of the probe over T₃ (for left temporal measurements) and over O₁ and O₂ (for occipital measurements) was guided by International 10-20 EEG electrode position classification system (b).

Figure 3

An individual infant’s data. The top left graph shows optical density changes from baseline in the occipital region over 300 seconds for 830 nm (HbO₂) and 690 nm (HbR). The bottom left graph shows optical density changes in the temporal region. Onset of each 20 s audiovisual trial is indicated by a red bar; onset of each 20 s visual trial is indicated by a pink bar. These alternate with 10 s no-stimulus baseline intervals (not marked). Graphs on the right reflect conversion of optical density to relative concentrations of oxygenated and deoxygenated hemoglobin, averaged across a maximum of four 60 s blocks (20 s audiovisual trial + 10 s baseline interval + 20 s visual trial + 10 s baseline interval). The top right graph shows the bilateral occipital response; the bottom right graph shows the left temporal response. The y-axes of both graphs indicate relative changes in concentration (micromolar) of HbO₂ and HbR.

Figure 4

Grand average hemodynamic (HbO₂, HbR, and HbT) response curves across 60-s blocks by cortical region. Audiovisual segments are indicated by the solid red bar (onset at time 0 s); visual segments are indicated by the solid pink bar (onset at time 30 s). Each trial was followed by a 10 s (no-stimulus) baseline interval. The y-axis indicates relative changes in concentration (micromolar) of the different chromophores.

Figure 5

Bar graph illustrates average change (and standard error) in HbO₂ concentration (micromolar) from time -2 to 0 seconds (baseline) to 10-20 seconds (second half of trial) for each stimulus condition and cortical region.