Objetivos: Estabelecer amostras de referência constituídas por gravações avaliadas com consenso como representativas da presença da fricativa faríngea (FF) e da plosiva faríngea (PF) e comparar as avaliações perceptivo-auditivas da presença da FF e da PF sem e com acesso às amostras de referência em três grupos com experiências distintas. Metodologia: O estudo foi dividido em duas etapas: durante a ETAPA 1, três fonoaudiólogas especialistas na identificação das alterações de fala associada com a fissura labiopalatina (FLP) ou a disfunção velofaríngea (DVF) analisaram um total de 136 amostras de fala com a tarefa de identificar o uso da FF, da PF ou do ponto articulatório correto durante produção dos sons [f], [s], [], [v], [z], [], [k] e [g]. As gravações foram avaliadas individualmente e aquelas que não apresentaram consenso inicial quanto à FF e PF foram novamente avaliadas de forma simultânea. As amostras avaliadas com 100% de concordância entre as 3 avaliadoras foram inseridas no Banco de Amostras Representativas da FF e da PF. Deste Banco de Amostras foram selecionadas um conjunto de 14 amostras para compor o material denominado Amostras de Referência e outro conjunto de 64 amostras denominadas Amostras Experimentais. A ETAPA 2 do estudo envolveu a análise das Amostras Experimentais por um grupo de 9 avaliadores sem e com acesso às Amostras de Referência as quais a condição com acesso foram usadas como recurso para calibrar os avaliadores. Os avaliadores foram agrupados de acordo com a experiência na identificação da FF e PF: G1 incluiu 3 discentes no primeiro ano do curso de fonoaudiologia; G2 incluiu 3 fonoaudiólogos pós-graduandos recém-formados e sem vivências em FLP e DVF, e G3 incluiu 3 fonoaudiólogos com experiência na avaliação de alterações de fala na FLP e DVF. Os avaliadores analisaram, individualmente, as amostras experimentais duas vezes, primeiro sem o acesso às Amostras de Referência e após um intervalo de 15 dias avaliaram as mesmas amostras com acesso controlado às Amostras de Referência. Foram estabelecidos: porcentagem de acerto quanto ao ponto articulatório usado (AC ou alvo) e valores de concordância intra e interavaliadores. Resultados foram analisados com ANOVA. Resultados: Na ETAPA 1 foram identificadas 49 amostras de ACs com concordância de 100%, sendo 42 representativas da FF e 7 representativas da PF. A concordância intra-avaliador nesta etapa foi considerada excelente para as três avaliadoras. Na ETAPA 2 foi verificado que os achados na condição com acesso às referências foram superiores aos achados sem acesso às referências. As diferenças em porcentagens de identificação correta da FF e PF entre as condições com e sem referências variaram entre o mínimo de 4% para G3 e máximo de 80% para G1, sugerindo performance consistentemente melhor para o grupo de avaliadoras com experiência na identificação de AC. Os valores médios de Kappa para os sons estudados indicaram que a concordância no G1 passou de regular (K=0,34) na condição sem referência para quase perfeita (K=0,97) com referência; no G2 passou de substancial (K=0,67) para quase perfeita (K=0,96); no G3 a concordância manteve-se quase perfeita nas duas situações (K=0,90 e K=0,91). Observou-se que sem o acesso às amostras de referência as melhores concordâncias ocorreram para o G3, seguido de G2, com as piores concordâncias para G1. Após o acesso às referências todos os grupos tiveram concordância quase perfeita. As diferenças entre os valores médios de Kappa nas duas condições foi significante. Conclusão: Os resultados sugerem que os valores de concordância e a porcentagem de acerto quanto ao uso de ponto articulatório correto e FF e PF para os alvos estudados aumentou significativamente com o acesso às Amostras de Referência. / Objectives: To establish reference samples representative of pharyngeal fricative (PF) and pharyngeal stop (PS) using consensual auditory-perceptual identification and to compare ratings of PF and PS with and without the use of the reference samples In three groups with different experiences. Methods: The study was divided into two stages: during the first stage, three speech-language pathologists (SLP) specialists in the identification of the speech disorders related with cleft lip and palate (CLP) and velopharyngeal dysfunction (VPD) analyzed 136 speech samples with to identify the use of PF, PS and the correct place of production of sons [f], [s], [], [v], [z], [], [k] e [g]. The recordings were rated individually and those without consensus regarding the PF and PS were rated again simultaneously. The samples rated with 100% agreement between the 3 SLPs were included into a Bank of Reference Samples Representative of PF and PS. A group of 14 and 64 samples were selected from the Bank creating the Reference Samples and Experimental Samples for stage 2. Stage 2 involved the rating of the Experimental Samples by a group of 9 evaluators without and with access to the Reference Samples to calibrate the raters. The raters were grouped according to the experience in the identification of PF and PS: G1 included 3 SLP students; G2 included 3 newly graduated SLPs and G3 included 3 SLPs experienced in rating PF and PS. The raters analyzed, individually, the Experimental Samples twice, first without access to the Reference Samples and 15 days later with controlled access to the Reference Samples. The percentage of correct ratings and the intra and interrater agreement were established. Findings were analyzed with ANOVA. Results: During Stage 1 49 samples representative of compensatory articulation (CA), 42 for PF and 7 for PS, were identified with 100% agreement between 3 experienced SLPs. Intrarater agreement during this stage was excellent for the 3 professionals. During Stage 2, all ratings obtained with access to the Reference Samples were superior to the ratings without access to the references. The differences in percentage of correct identification of PF and PS between the conditions with and without access to the references varied between the minimum of 4% for G3 to maximum of 80% for G1, suggesting a consistently better performance for the group with raters experienced with identification of CA. The mean Kappa scores for all sounds indicated that the interrater agreement for G1 improved from regular (K=0,34) in the condition without references to almost perfect when the references were used, for G2 improved from substantial (K=0,67) to almost perfect (K=0,96), while for G3 the agreement was almost perfect in both conditions (K=0,90 e K=0,91). Without the access to the references better agreements were observed for G3, followed by G2 with worse agreement for G1. After access to the references all groups presented with almost perfect agreements. The differences in mean Kappa scores between the two conditions were significant. Conclusion: The findings suggest that raters agreement and percentage of correct identification of place of production and PF and PS for all sounds studied increased significantly with the access to Reference Samples.

Objetivos: Estabelecer amostras de referência constituídas por gravações avaliadas com consenso como representativas da presença da fricativa faríngea (FF) e da plosiva faríngea (PF) e comparar as avaliações perceptivo-auditivas da presença da FF e da PF sem e com acesso às amostras de referência em três grupos com experiências distintas. Metodologia: O estudo foi dividido em duas etapas: durante a ETAPA 1, três fonoaudiólogas especialistas na identificação das alterações de fala associada com a fissura labiopalatina (FLP) ou a disfunção velofaríngea (DVF) analisaram um total de 136 amostras de fala com a tarefa de identificar o uso da FF, da PF ou do ponto articulatório correto durante produção dos sons [f], [s], [], [v], [z], [], [k] e [g]. As gravações foram avaliadas individualmente e aquelas que não apresentaram consenso inicial quanto à FF e PF foram novamente avaliadas de forma simultânea. As amostras avaliadas com 100% de concordância entre as 3 avaliadoras foram inseridas no Banco de Amostras Representativas da FF e da PF. Deste Banco de Amostras foram selecionadas um conjunto de 14 amostras para compor o material denominado Amostras de Referência e outro conjunto de 64 amostras denominadas Amostras Experimentais. A ETAPA 2 do estudo envolveu a análise das Amostras Experimentais por um grupo de 9 avaliadores sem e com acesso às Amostras de Referência as quais a condição com acesso foram usadas como recurso para calibrar os avaliadores. Os avaliadores foram agrupados de acordo com a experiência na identificação da FF e PF: G1 incluiu 3 discentes no primeiro ano do curso de fonoaudiologia; G2 incluiu 3 fonoaudiólogos pós-graduandos recém-formados e sem vivências em FLP e DVF, e G3 incluiu 3 fonoaudiólogos com experiência na avaliação de alterações de fala na FLP e DVF. Os avaliadores analisaram, individualmente, as amostras experimentais duas vezes, primeiro sem o acesso às Amostras de Referência e após um intervalo de 15 dias avaliaram as mesmas amostras com acesso controlado às Amostras de Referência. Foram estabelecidos: porcentagem de acerto quanto ao ponto articulatório usado (AC ou alvo) e valores de concordância intra e interavaliadores. Resultados foram analisados com ANOVA. Resultados: Na ETAPA 1 foram identificadas 49 amostras de ACs com concordância de 100%, sendo 42 representativas da FF e 7 representativas da PF. A concordância intra-avaliador nesta etapa foi considerada excelente para as três avaliadoras. Na ETAPA 2 foi verificado que os achados na condição com acesso às referências foram superiores aos achados sem acesso às referências. As diferenças em porcentagens de identificação correta da FF e PF entre as condições com e sem referências variaram entre o mínimo de 4% para G3 e máximo de 80% para G1, sugerindo performance consistentemente melhor para o grupo de avaliadoras com experiência na identificação de AC. Os valores médios de Kappa para os sons estudados indicaram que a concordância no G1 passou de regular (K=0,34) na condição sem referência para quase perfeita (K=0,97) com referência; no G2 passou de substancial (K=0,67) para quase perfeita (K=0,96); no G3 a concordância manteve-se quase perfeita nas duas situações (K=0,90 e K=0,91). Observou-se que sem o acesso às amostras de referência as melhores concordâncias ocorreram para o G3, seguido de G2, com as piores concordâncias para G1. Após o acesso às referências todos os grupos tiveram concordância quase perfeita. As diferenças entre os valores médios de Kappa nas duas condições foi significante. Conclusão: Os resultados sugerem que os valores de concordância e a porcentagem de acerto quanto ao uso de ponto articulatório correto e FF e PF para os alvos estudados aumentou significativamente com o acesso às Amostras de Referência. / Objectives: To establish reference samples representative of pharyngeal fricative (PF) and pharyngeal stop (PS) using consensual auditory-perceptual identification and to compare ratings of PF and PS with and without the use of the reference samples In three groups with different experiences. Methods: The study was divided into two stages: during the first stage, three speech-language pathologists (SLP) specialists in the identification of the speech disorders related with cleft lip and palate (CLP) and velopharyngeal dysfunction (VPD) analyzed 136 speech samples with to identify the use of PF, PS and the correct place of production of sons [f], [s], [], [v], [z], [], [k] e [g]. The recordings were rated individually and those without consensus regarding the PF and PS were rated again simultaneously. The samples rated with 100% agreement between the 3 SLPs were included into a Bank of Reference Samples Representative of PF and PS. A group of 14 and 64 samples were selected from the Bank creating the Reference Samples and Experimental Samples for stage 2. Stage 2 involved the rating of the Experimental Samples by a group of 9 evaluators without and with access to the Reference Samples to calibrate the raters. The raters were grouped according to the experience in the identification of PF and PS: G1 included 3 SLP students; G2 included 3 newly graduated SLPs and G3 included 3 SLPs experienced in rating PF and PS. The raters analyzed, individually, the Experimental Samples twice, first without access to the Reference Samples and 15 days later with controlled access to the Reference Samples. The percentage of correct ratings and the intra and interrater agreement were established. Findings were analyzed with ANOVA. Results: During Stage 1 49 samples representative of compensatory articulation (CA), 42 for PF and 7 for PS, were identified with 100% agreement between 3 experienced SLPs. Intrarater agreement during this stage was excellent for the 3 professionals. During Stage 2, all ratings obtained with access to the Reference Samples were superior to the ratings without access to the references. The differences in percentage of correct identification of PF and PS between the conditions with and without access to the references varied between the minimum of 4% for G3 to maximum of 80% for G1, suggesting a consistently better performance for the group with raters experienced with identification of CA. The mean Kappa scores for all sounds indicated that the interrater agreement for G1 improved from regular (K=0,34) in the condition without references to almost perfect when the references were used, for G2 improved from substantial (K=0,67) to almost perfect (K=0,96), while for G3 the agreement was almost perfect in both conditions (K=0,90 e K=0,91). Without the access to the references better agreements were observed for G3, followed by G2 with worse agreement for G1. After access to the references all groups presented with almost perfect agreements. The differences in mean Kappa scores between the two conditions were significant. Conclusion: The findings suggest that raters agreement and percentage of correct identification of place of production and PF and PS for all sounds studied increased significantly with the access to Reference Samples.

Reference samples are a vital part of the forensic analysis of deoxyribonucleic acid (DNA) evidence. Efficient processing and analysis of these sample types are required for comparative analysis of an unknown electropherogram (EPG) and forensic databasing purposes (12). These reference samples can be derived from blood swabs or cheek swabs, the latter also being known as buccal cell swabs (20, 22, 32). Buccal cells, or epithelial cells of the oral cavity, are the preferred cell type for known samples as their collection is non-invasive and painless (20-21). Buccal cell collection devices typically consist of a swab (cotton or foam) and a filter paper, commonly FTA paper (1). FTA paper contains proprietary chemicals that lyse cell membranes upon contact, trapping and stabilizing DNA for downstream processing (21, 34). FTA paper also inhibits bacterial and viral growth and protects against damage from UV radiation, nucleases and oxidation (21, 34). Some of the benefits of using FTA cards include the ability to store the cards at ambient temperature for years (21, 35, 37) and to perform direct amplification of the samples thereby removing the need to utilize DNA extraction and quantitative polymerase chain reaction (qPCR) (32, 37, 39). The EasiCollectTM (EC) and EasiCollectTM + (EC+) Buccal Sample Collection Devices (General Electric (GE) Healthcare Life Sciences, Buckinghamshire, UK) have FTA sample collection cards that contain a proprietary dye that changes color from pink to white, indicating where colorless fluids, such as saliva, were likely deposited (42). This study consisted of four phases. Phase 0 determined the optimal amplification conditions, including number of polymerase chain reaction (PCR) cycles and an appropriate capillary electrophoresis (CE) injection time for high template, single source samples obtained from FTA cards using the EC and EC+ buccal cell collection devices. Samples were obtained from EC FTA cards with a Harris 1.2-mm Uni-Core Punch and amplified using the GlobalfilerTM Express PCR Amplification Kit (Thermo Fisher Scientific, Waltham, MA) using the manufacturer’s protocol with 26, 27 or 28 PCR cycles (28). Fragment separation was achieved on an ABI 3500 Genetic Analyzer (Applied Biosystems, Foster City, CA) with 5, 15 and 25 second (s) 1.2 kiloVolt (kV) injections. Samples were analyzed on GeneMapper® ID-X v1.4 (Applied Biosystems, Foster City, CA) with an analytical threshold of 150 RFU (relative fluorescence units) (31). It was determined that amplification with 26 PCR cycles was optimal for high template, single source samples from FTA cards in this laboratory. The three injection times were utilized in the remaining phases and no other parameters were changed. In Phase 1 of this study, the optimal collection method for the EC+ device from various processes was assessed using the following collection variables: 1) a dry or saliva-wet swab; 2) a circular or up-down/side-to-side motion; 3) 2, 3 or 4 motions; and 4) swabbing of one or both cheeks. This resulted in a total of 24 distinct collection processes. We found that collection techniques that involved wetting the foam head of the EC+ device provided higher peak heights, improved heterozygote balance (Hb) and minimized the rate of drop-out in EPGs. When swabbing two cheeks versus one, the median peak heights increased, indicating an increase in transfer of cellular material onto the FTA surface. The motion of swabbing - circular or up-down/side-to-side - did not have an effect on the overall quality of the EPG data. During Phase 2a, the distribution of cellular material was assessed for two collection processes that involved swabbing of two cheeks with a wet swab four times; the variation among the methods being the motion (circular or up-down/side-to-side). Two punches taken surrounding the original punch assessed during Phase 1 showed similar average peak heights (i.e. ca. 3500 RFU at a 5 s injection on the ABI 3500 Genetic Analyzer) for both collection processes. No allelic drop-out was observed with either collection technique. Phase 2b compared the EPG signal of the EC and EC+ collection devices. The EC+ collection process used for this comparison involved rubbing a wet swab across two cheeks using four circular motions as this produced no allelic drop-out and fewer samples which saturated the CE laser detector. Therefore, this method provided more data for analysis. Samples from both devices produced comparable peak heights and PHRs above 0.6 with no allelic drop-out and stutter ratios below the thresholds set by the manufacturer (28). The EC+ device was found to be robust and provided full profiles using a minimalist sample collection method. However, the probability of drop-out increased as both the number of motions and the number of cheeks decreased. Based on this study, a collection using four circular motions divided between two cheeks with a wet swab is recommended with a 5 s, 1.2 kV injection on an ABI 3500 Genetic Analyzer, since full DNA profiles were obtained with balanced heterozygote loci, expected stutter ratios, and acceptable levels of minus A artifact. Further, it was determined that this recommended collection method resulted in high-fidelity DNA signal for up to three punches. Thus, the EC+ device is reliable, easy-to-use and non-invasive for the collection of buccal cells for known reference samples. A sample obtained from the area of transfer on an FTA card from the EC+ device can produce an EPG of the quality required for the comparison of known samples to an evidentiary profile as well as for input of the genotype into a national forensic database.