Babylab

Introduction

Parents sing simple songs and chant rhythmic rhymes to babies to sooth or to amuse them well before they can sing or talk. We all intuitively feel that babies like rhythms and melodies even though we don’t know how they perceive them. In fact, for a long time the prevailing view was that babies are born into a chaotic perceptual world of light and sound. However, in the past 20 years, careful observations and investigations have led scientists to believe that young infants process sensory input to a much higher degree than previously thought. In search of the origin and development of music perception, our aim is to test in newborn babies the operation of some of the simplest processing steps required for perceiving music. Within the EMCAP project, we are investigating questions such as whether or not newborns possess the ability to form sound groups, or to process pitch intervals independently of the absolute pitch level.

How can we tell what newborn babies hear?

Studying sound processing in newborn infants is difficult, because of course they cannot be asked about what they heard and in general, they do not show behavioural responses, which could be reliably linked to auditory processes. Moreover, most physiological measures used in adults for assessing brain processes may either place newborns at risk or disturb their normal daily rhythms; e.g., wake them up. One exception is the recording of electrical brain activity, which is safe and can be conducted even while they are asleep. Electric brain responses elicited by sounds can tell us about the processing of auditory information in the brain even when the subject is not aware of the sounds, such as in sleep. For studying music perception in newborn babies, we have established a laboratory in the maternity ward of Clinic I for Obstetrics and Gynaecology of the Semmelweis University in Budapest, Hungary. Here we will measure electric brain waves in neonates in response to sound patterns. 

Ground breaking research

Electric brain activity (electroencephalography, abbreviated as EEG) is measured with electrodes attached to skin on various points of the scalp, which are connected to amplifiers and signal recording equipment. Although EEG measurement is often used for medical and scientific examinations in adults, recording EEG signals in newborns requires even stricter selection of materials and procedures. Sound delivery is also a difficult issue, because there are few types of headphone that can be used for delivering sounds to sleeping newborns. Finally, EEG signals in newborns are much more variable than in adults and therefore, a larger number of measurements and careful data processing are required for successfully interpreting the signals. Due to these and other difficulties, only a few laboratories around the world conduct EEG studies in newborns for research purposes. Thus the newborn EEG laboratory established within the EMCAP project will provide novel and important information for understanding the origins of music perception.

Video

Early Results

Newborn Hearing Screening with Brainstem Auditory Evoked Potentials

Averaged Evoked Brain Responses

Rare noise stimuli presented amongst frequent tones activate a number of processes in human auditory system. The early upward going deflection, best seen on the frontal electrodes (F3 - left frontal and F4 - right frontal), represents the response to change in the spectral make-up of the sound (the noise contains frequencies that were not present in the frequent tones). The following downward-going deflections, best seen over the central scalp - Cz, reflect neuronal activity elicited by the higher intensity of the noise compared with the frequent tones. Finally the small late upward-going deflection is probably a sign showing that babies form crude sound categories (simple tones vs. complex noise sounds). An unexpected categorical change in the sound input indicates the arrival of new information that requires further processing.

Note how much information the auditory system of newborns extracted from the noise sound within around half a second!

Higher-level Auditory Function: Grouping by Pattern Repetition (Preliminary Results)

Figure 1.

 Our first study is almost complete with electrical brain responses having been measured in 10 newborn babies. The results so far suggest that newborn infants detect repeating pitch patterns similarly to adults.

The figure shows auditory event‑related brain potentials (ERP), changes in the electrical brain activity time‑locked to the delivery of some sound. In this case, electric brain responses were separately averaged for two different tones, “A” and “B”, which differed from each other only in frequency. The onset of the tones is at the crossing of the x and y axes with the x axis representing time in milliseconds and the y axis showing the voltage in mV. The ERPs shown have been filtered between 2.5 and 20 Hz.

Two stimulus conditions were administered. In the “Random Condition”, the “A” and “B” tones (termed “Standard” and “Deviant”, respectively) were delivered in a randomized order at a 10 Hz stimulus delivery rate. The “B” tones appeared less frequently (20% of the time) the “A” tones (80%). Because “B” tones broke the repetition of the “A” tone, they were expected to be processed differently from “A” tones. The latency range of the expected ERP difference is highlighted on the figure. In the "Grouped condition", a repeating AAAAB pattern was presented. Note that the overall presentation frequency of “B” tone within the stimulus sequence is the same as in the “Random Condition” (20%). If the repetition of the AAAAB pattern was detected, then the response elicited by the “A” and “B” tones should not differ from each other. This is because the AAAAB pattern then becomes the unit of perception and thus the “B” tone does not break the emerging higher-order regularity. From these preliminary data it appears that, similarly to adults, newborns can perceive repeating temporal patterns.

Figure 2.

Wavelet-based time-frequency analysis of single-trial data revealed prominent peaks of electrical brain activity in the gamma-band (~30-40 Hz). These results showed that the electrical brain activity evoked by the “B” tones (“Deviant”, left panels) were processed differently in the “Random Condition” (upper panels) than in the “Grouped Condition” (lower panels).

The ERSP (Event-Related Spectral Perturbation) matrices show event-related changes in EEG power as a function of frequency (the y axis calibrated in Hz) and the time from the onset of tones (the x axis calibrated in ms, 0 representing the onset of the tone). Above‑average activity is marked by warm colors, below‑average by cold colors.

In the “Random Condition”, „B” tones evoked strong gamma‑band oscillations at ~200 ms poststimulus with a center frequency of ~36 Hz (top left), whereas standard „A” tones (“Standard”) were followed by a decrease of gamma‑band brain activity (top right). These preliminary findings suggest that, similarly to adults, newborn infants show gamma oscillatory activity, which is sensitive to stimulus probability. The time-frequency profiles of the brain activity evoked by „B” and „A” tones in the "Grouped condition" (Figure 2., bottom) are different from those observed for the “Random Condition”. One may detect a pattern of activity that repeats with the cycle of the stimulation. The pattern of increased activity between 100 and 200 ms from the onset of the “A” tone preceding the “B” tone by 3 positions (lower right panel) appears to be very similar to the pattern following the onset of the “B” tone by 300‑400 ms (lower left panel). The latency differences add up to a full cycle of the tone pattern (500 ms). Thus this brain activity may play role in the formation of auditory perceptual units in newborns.