*2.3. Measures*

#### 2.3.1. Outcome Measures: Coding Scheme for Communicative Interaction

A coding scheme was used to investigate the characteristics of communicative interaction between the dyads, based on previous video-coding research on children with multiple disabilities and complex communication needs [5,14,15,27] with adaptations to fit the EGAT condition. The coding scheme applied to code all behaviors within the time frame of interaction included interactional structure, communicative functions, and modes of communication.

Interactional structure was classified as turns and moves [5,14]. Turns were determined by a succession of communicative signs with the boundary between turns a twosecond gap supported by the presence of other behaviors, for example, non-verbal signals, pitch change, the listener took a turn, or the speaker came to a rest [14,15,27]. Each turn could include one or more moves. A move as defined by Pennington and McConachie comprised "single or strings of utterances/non-verbal communicative signals produced by one speaker within a conversational turn" [15], p.398. Moves included Initiation (I), opening the conversation or introducing a topic and could solicit a response; Response (R), a reply to an initiation; Response/Initiation (R/I), a reply to initiation but also requiring a response of its own; Follow-up (F), acknowledging the previous utterance; Follow-up/Initiation (F/I), acknowledging the previous move and requesting a response of its own; No Response. Table 1 displayed an example.


**Table 1.** Example of moves in a play activity.

In this study, Preparatory (P) and Operation/Navigation (ON) were added to interactional moves in consideration of the preparation act to make ready for interaction (P) [5] and when the participants spend time struggling with low-tech AAC systems or EGAT with computers (ON) [23].

Communicative functions were coded to represent the intentions and purpose of the communicative act [14,15]. Each move could contain one or more codes of communicative functions [14] if the communicative purposes in the context occurred simultaneously. The categories included Requestive (RE), request for attention, information, action or clarification; Informative (IN), comments, answers, or clarification of a previous utterance; Acknowledgment (ACK), conveying understanding to previous utterance; Confirmation/Denial (CD), affirmation or rejection; Self/shared expression (SSE), demonstration of the emotional state in individuals; and Unintelligible (U), not understandable by a listener or a coder. An example was presented in Table 2.

**Table 2.** Example of communicative functions in a play activity.


Note. RE = Requestive; IN = Informative; SSE = Self/shared expression.

Modes of communication were defined as the means by which communicative functions were transmitted, including speech, vocalization, gesture, and aided AAC systems [5,14,15]. The modes were combined and coded if they appeared to signal the same communicative function. Low-tech mode (Lt) was used to indicate the individual communicated using low-tech AAC, e.g., a communication book or Bliss board, and EGAT was coded to represent that persons used eye-gaze technology for aided communication.

All categories and definitions of the codes are presented in the Appendix A Table A1.

#### 2.3.2. Measures Related to Participant Characteristics

This study received participant information from the medical charts by research coordinators in the organizations. Motor severity levels were based on Gross Motor Function Classification System (GMFCS) with level I (ambulatory without restrictions) to V (limited ability to move around) [30], and Manual Ability Classification System (MACS), with level I (handles objects easily and successfully) to V (does not handle objects) [31]. Medical diagnosis, sensory functions (e.g., vision), severity of motor impairments and cognitive impairments of the children/youths were documented.

In addition, eye-control skills and communication abilities of the children/youths were assessed by Compass Aim test [32] and Communication Matrix [33], respectively, to describe the essential abilities related to use of EGAT. A trained AT specialist in the participating organization performed assessments before the EGAT intervention started.

#### Compass Aim Test

The Compass Aim test measures eye-gaze performance in computer interaction, encompassing two variables: (1) Accuracy, to measure the ability to control a mouse pointer for target selection, and (2) Time on task, to measure the required time for target selection on the screen. The test showed high test–retest reliability, good internal consistency and adequate construct validity [32].
