A New Approach for Eliciting Expressive Language Samples: Elsa

Background: Emergence of expressive language is one of the strongest predictors of positive outcomes in individuals with ASD (Howlin et al., 2004; Venter et al., 1992). Language abilities also play an important role in social communication, which places them at the center of numerous interventions, and consequently requires the use of reliable and valid language measures for their assessment. Natural language samples collected across variety of contexts provide great measures of speakers’ expressive language (Tager-Flusberg et al., 2009). Yet, no protocol for language elicitation has been specifically designed for individuals with ASD across a range of ages and language abilities. This is particularly important considering that language ability is one of the most variable characteristics of diagnosed individuals with 30% failing to acquire functional speech (Kjelgaard & Tager-Flusberg, 2001). Many studies rely on the ADOS as a language sampling context. However, presenting participants with different presses and activities depending on the module administered might influence language measures.

Objectives: We aim to evaluate a new Eliciting Language Samples Assessment (ELSA) protocol for individuals with ASD across a range of ages and language ability. The study focuses on its use with minimally verbal children with ASD, arguably the most challenging group from whom to obtain language samples. We evaluated the effectiveness of ELSA in eliciting language and engaging the participants. We compared it to a different language sampling context, the ADOS.

Methods: ELSA includes multiple activities designed to engage individuals across a wide age range including going on a hike, camping, doing crafts (all carried out using a kit with relevant props and materials), watching a short video and retelling its plot, and a short conversation period. We evaluated protocol engagement by examining the duration of engagement of the participants (how long it took to go through all the activities). Utterance frequency per minute was used to assess the effectiveness of ELSA in eliciting language, and compare it to the ADOS.

Results: Participants (N=20) aged 6;7 to 18;10 spent on average 18.96 (SD=4.49) minutes engaging with ELSA. They produced an average of 3.79 (SD=3.91) utterances per minute. The frequency of utterances per minute during ELSA (M=3.1817) and ADOS (M=2.2760) approached a significant difference (t(8)= 2.0491, p=0.0746) with more utterances per minute during ELSA.

Conclusions: ELSA was successful at engaging participants across a wide age range and language ability for a prolonged period of time (Table-1). In addition, participants, even minimally verbal ones, produced speech sounds while engaging with the ELSA activities. The video component of the protocol was originally included in order to evaluate the narrative skills of more verbally fluent participants, so it might not be applicable to less verbally fluent individuals as reflected by the small number of vocalizations they produced during this portion of ELSA. Participants tended to make utterances more frequently during ELSA than during the ADOS, which speaks to ELSA’s ability to elicit language. These analyses provide support for the use of the newly designed ELSA for the assessment of the highly heterogeneous language ability in ASD.