This protocol presents a novel methodology for the neural decoding of intent from freely-behaving infants during unscripted social interaction with an actor. Neural activity is acquired using non-invasive high-density active scalp electroencephalography (EEG). Kinematic data is collected with inertial measurement units and supplemented with synchronized video recording.
Understanding typical and atypical development remains one of the fundamental questions in developmental human neuroscience. Traditionally, experimental paradigms and analysis tools have been limited to constrained laboratory tasks and contexts due to technical limitations imposed by the available set of measuring and analysis techniques and the age of the subjects. These limitations severely limit the study of developmental neural dynamics and associated neural networks engaged in cognition, perception and action in infants performing “in action and in context”. This protocol presents a novel approach to study infants and young children as they freely organize their own behavior, and its consequences in a complex, partly unpredictable and highly dynamic environment. The proposed methodology integrates synchronized high-density active scalp electroencephalography (EEG), inertial measurement units (IMUs), video recording and behavioral analysis to capture brain activity and movement non-invasively in freely-behaving infants. This setup allows for the study of neural network dynamics in the developing brain, in action and context, as these networks are recruited during goal-oriented, exploration and social interaction tasks.
One of the fundamental human capabilities is the ability to learn and deploy actions (action production) strategically in service of goals and rewards, the ability to apprehend the goals of social partners (action understanding) in order to produce appropriate social responses, and the ability to learn from others through observation and imitation 1. The neural basis of these cognitive-motor capabilities have been attributed, at least in part, to the so-called mirror neuron system; a system that is thought to be engaged when one views someone performing an action and when one performs the action. However, the potential link between the mirror neuron system and action understanding is not yet well understood 1. Studying the emergence and development of this mirror neuron system in human infants has been hampered by a) the technical limitations of multi-modal data acquisition of brain activity correlated to intent and fine-grained motion data, b) the constraints imposed by experimental protocols that are unnatural (e.g., social interaction with an agent depicted in a videotape, the need for maintaining a still posture to minimize artifacts during Electroencephalographic (EEG) recordings, etc.), and c) the communication/language barriers when testing young infants/toddlers that greatly limit the researcher’s capabilities to give instructions and validate behaviors.
For a better understanding of the varying neural and behavioral dynamics in natural behavior, we developed a novel experimental and analytical approach that allows the time-resolved study of the neural substrates of emerging goal-oriented and social behaviors in young children. Specifically, we deployed an EEG based mobile brain imaging (MoBI) approach 2 to record brain activity and movement from freely-behaving infants during interaction with an experimenter. Inertial measurement units (IMUs) were used to monitor subject and experimenter’s kinematics.
EEG technology and inertial sensors were used to study neural patterns and activations associated to the infants’ action imitation and goal-oriented behaviors in an unscripted interaction with an experimenter/actor. Actions such as reach-grasp, reach-offer, observe, rest, and explore are all part of the cognitive-motor processes involved in imitation. Furthermore, we use source estimation to localize the generators of electric potentials within the brain during the behavioral tasks, thereby studying the spatiotemporal dynamics of neuronal currents throughout the brain. Similarly we deploy machine learning algorithms to assess and measure the predictability of these behavioral actions by identifying action-relevant spatio-temporal patterns in the neural activity in sensor (EEG) and/or source spaces. Integrating traditional ERD/ERS, source and decoding analysis provide a more comprehensive developmental description of the neural basis of such behaviors.
This setup allowed us to exploit the advantages of the MoBI approach 2,3 and study the social interactions between the infant and the experimenter as they naturally occur without restrictions.
The protocol, from the time the subject arrives to the time he/she leaves, takes approximately 1 hr to complete. The IMU/EEG setup time and electrode location acquisition varies from 15 – 25 min depending on factors such as hair length and cooperation of the subject. The initialization and configuration of the equipment adds up to 10 min, and the testing session lasts approximately 15 min. Removal of the IMUs and EEG cap, including cleaning the head of the infant from the hypoallergenic gel, takes 5-10 min.
The protocol described presents a methodology to collect data from freely-behaving infants while they are interacting with an experimenter in real time. It employs mobile brain imaging technology (scalp EEG) to capture neural activity while simultaneously recording kinematic data with IMUs in strategic body locations. The experiment session is also recorded by a video camera. The three data recording systems are synchronized using a custom trigger system.
The EEG and IMU system are strapped to the subject as he/she is freely moving during the experiment session. The IMUs need to be strapped securely to be able to capture kinematics accurately. To ensure full unconstrained mobility of the subject, the equipment has to be as minimally intrusive as possible; thus the use of the holder to support the EEG’s electrode cables and the EEG control box. The experimenter then interacts with the infant for approximately 15 minutes. The infant will elicit an age-dependent repertoire of behaviors during the interaction. These include rest, reach-grasp, reach-offer, explore, observe, and imitate. However, some infants will be unwilling to cooperate in the session due to fatigue, lack of comfort, or stress. Make sure to schedule the experiment when the child is most dynamic and active to prevent the occurence of negative responses from him/her.
The nature of the experiment presents risks to the quality of the data recorded throughout the session. Therefore, it is crucial to test all the connections and data quality before starting the recording session, and to monitor them continuously during the session. If the data EEG system is not recording quality data, stop the software and unplug all connections. Before restarting the software or connecting the equipment back to the laptop, remove all possible noise sources (i.e., power supplies) from the proximity of the recording hardware. The EEG hardware includes signal amplifiers that can pick up environmental noise if placed close to electrical noise sources. For the IMU receiver, make sure there is no interference in the line of sight between the receiver and the experiment and infant.
This experimental setup provides high temporal resolution neural data by measuring electrical activity at the surface of the scalp. Recent studies have demonstrated the feasibility of utilizing these signals, together with whole-body kinematics, to identify classifiable information for expressive movements 20, and functional movements 21,22, suggesting that this proposed data collection approach could lead to a better understanding of the neural basis of imitation in infants.
Recent contributions featuring powerful machine-learning algorithms applied to brain dynamics 13,20,21 are building a growing toolkit to study surface potentials in more natural settings. This proposed setup provides a spectrum of possibilities for research questions to be addressed 2,22. Particularly, it can be applied to research focused on a) understanding the neural basis of cognitive-motor development of infants based on a large population of subjects; b) understanding the neural basis of the baby’s intent in ‘action and context’, which should be predictive of the incoming behavioral action; c) quantifying common and unique neural patterns to characterize individuality and variability in the developing brain; and d) studying the emergence of imitation and learning processes. These goals entail the deployment of machine learning algorithms that can deal with statistically rich data both in informative brain-originated potentials and movement or muscle artifacts 12,20,23 .
This study attempts to estimate the cortical sources and electric field potentials using the infant EEG data. Due to the technical difficulties such as the lack of knowledge in infant head conductivity values and thickness of cortical matter, the accurate modeling of the head model is a difficult task. Further studies are needed for noninvasive regional tissue conductivity estimates in infants 24. Cortical surface segmentation of infant MRI data presents an additional challenge due to the poor contrast found in images of the developing human brain 25. Future research is needed to address these difficulties and estimate various neurophysiological correlates of infant development and behavior.
Finally, the proposed experimental protocol and methods could be deployed in the study of those with developmental disabilities such as infants with probable autism spectrum disorder (ASD). In such an application, it would be desirable to include a control group and suitable developmental assessments to characterize the two groups (control and ASD). For example, a study group could consist of all high-risk (for ASD) infant siblings assessed with the Autism Diagnostic Observation Schedule 26, symptom severity 27 and the Mullen Scales of Early Learning 28 to characterize general cognitive ability. If available, diffusion-weighted MRI scans would also be highly desirable 29.
The authors have nothing to disclose.
This work was supported by Eunice Kennedy Shriver National Institutes of Child Health & Human Development (NICHD) Award # P01 HD064653-01. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NICHD or the National Institutes of Health.
BrainAmp Amplifier | Brain Products, Gmbh | ||||
actiCAP EEG cap- 64 electrodes* | Brain Products, Gmbh | ||||
actiCAP Control Box | Brain Products, Gmbh | ||||
Brain Vision Recorder software v1.20.0601 | Brain Products, Gmbh | ||||
actiCAP Control Box software v1.2.5.2 | Brain Products, Gmbh | ||||
CapTrak | Brain Products, Gmbh | ||||
CapTrak software v1.0.0 | Brain Products, Gmbh | ||||
Opal movement monitor | APDM, Inc | ||||
Opal docking station | APDM, Inc | ||||
Opal wireless access point | APDM, Inc | ||||
Motion Studio software v1.0.0.201503302222 | APDM, Inc | ||||
Trigger box | Custom | ||||
Video camera | HC-W850M, Panasonic Co. | ||||
*The EEG caps come in the following head circumference sizes for infants: 42, 44, 46, 48, 50cm. For this protocol, a stock of 2 caps of each size is recommended. |