METHOD FOR EARLY DIAGNOSIS OF AUTISM SPECTRUM DISORDER IN CHILDREN

Abstract

A method for early diagnosis of autism spectrum disorder (ASD) in a child, characterized in that four movement sensitivity parameters are measured during an interaction of examined child with a touch screen, characterized in that: Linear Movement Precision (LMP) is measured by calculating variation between a line drawn by ASD child and a line drawn by TD (typically developing) child; Polygonal Movement Precision (PMP) is measured by calculating the smallest area of a polygon drawn with knots generated by ASD child and of a polygon drawn with knots generated by TD child; Movement Acceleration Stability (MAS) is measured by calculating jerk defined as first derivative of acceleration while drawing a line by ASD child and while drawing a line by TD child; Pressing Force Value (PFV) is measured so that the touch screen is placed on flat surface and force applied by ASD child or TD child while drawing a line is measured, and wherein, the force values are calculated and compared.

Claims

1. A method for early diagnosis of autism spectrum disorder (ASD) in a child, characterized in that four movement sensitivity parameters are measured during an interaction of examined child with a touch screen, wherein: i) Linear Movement Precision (LMP) is measured by calculating variation between a line drawn by ASD child and a line drawn by TD (typically developing) child, and wherein the numbers of knots used to draw both lines are calculated and compared; ii) Polygonal Movement Precision (PMP) is measured by calculating the smallest area of a polygon drawn with knots generated by ASD child and of a polygon drawn with knots generated by TD child, and wherein these knots are connected with line sections and the smallest areas of resulting polygons are calculated and compared, iii) Movement Acceleration Stability (MAS) is measured by calculating jerk defined as first derivative of acceleration while drawing a line by ASD child and while drawing a line by TD child, and wherein the jerk values are calculated and compared; iv) Pressing Force Value (PFV) is measured so that the touch screen is placed on flat surface and force applied by ASD child or TD child while drawing a line is measured, and wherein the force values are calculated and compared.

2. A method according to claim 1, characterized in that a mobile device is applied and in that the touch screen is the one of a mobile device.

3. A method according to claim 2, characterized in that the mobile device is a tablet or a smartphone.

4. A method according to claim 1, characterized in that it comprises a procedure of assessing a sharing ability of the child which is distributing food between four cartoon characters on the touch screen, and wherein the procedure consists of a series of sharing trials, where every new trial means distribution of new portion of food, and wherein the child's task is identical in each trial, however the food item, which the child has to divide and share, differs from trial to trial, and wherein food can be easily divided into four even rations with a simple gesture, each ration can then be dragged and dropped onto the plate in front of each cartoon character which then makes a positive feedback to the child if provided with the food ration.

5. A method according to claim 1, characterized in that it comprises a procedure of assessing a creativity ability of a child which is instructed to outline and colour a shape on the touch screen, and wherein the child first chooses the shape he/she wants to draw, then outlines it step-by-step with his/her finger, and finally colours it using a colour palette, and wherein during the procedure precision and stability of the movements are measured.

6. A method according to claim 1, characterized in that the visual and/or audio screen are supplemented with one or more distractors which are moving, emitting sounds and/or are interactive in course of the procedure, and thus interfere with what is displayed on the touch screen and/or made audible to examined child.

Description

[0023] The proposed method is illustrated on the drawing, where:

[0024] FIG. 1 shows general idea of the methodology;

[0025] FIG. 2 shows first exemplary heat map for the game Sharingfor even sharing;

[0026] FIG. 3 shows second exemplary heat map for the game Sharingfor uneven sharing;

[0027] FIG. 4 shows positive feedback if the food has been shared evenly;

[0028] FIG. 5 shows exemplary screen in Creativity game; and

[0029] FIG. 6 shows a diagram of the modeling algorithm.

[0030] The machine learning approach is used to process data obtained from the tablet sensors as well as basic information about the child (age, sex). The model calculates the probability that a particular child's data belong to the ASD group or the control one (TD).

[0031] Parameters: 272 features were extracted from the device data to give a comprehensive computational description of the child's movements sensed by the device, and made in interaction with it. These features were obtained from the Screen (108 features) and Inertial (164 features) data. Of these features, 26 were highly correlated (r>0.9), considered redundant, and reduced to a single feature. 247 features from device sensors and touch, together with information about child's age and gender, were included in the final analysis.

[0032] Goal function: The fitting of the machine learning algorithm is conducted by solving an optimisation problem, which is to minimise the error between the clinical diagnosis (ASD or not) and the results obtained with the proposed method (probability of having autism) by changing fitting parameters.

[0033] Building the model: To build models able to predict group classification (ASD or control TD), and to ensure reliable classification, a k-fold cross validation procedure was employed. This method was used to establish the predictive power of the model, i.e. how the result ought to generalise to an independent dataset. To increase the stability of the result, additional k repetitions of the process were performed. The full dataset of calculated features was split into k equal sized samples. From k subsamples one was chosen for the validation (test), and the rest (k1) were used as the training dataset. Every sample was used exactly once as a validation (test) dataset. This process was performed k times (folds). During every iteration model was trained on the k1 sample and then tested on the one sample left for the prediction. Results of the prediction were stored, to later establish the end-point model, which combines the results from each fold. This process was repeated k times. In the end kk samples were created and tested. Based on the prediction data gathered during both iterations (repetitions and folds), the Receiver Operating Characteristics (ROC) curve was generated, and the sensitivity (true positive rate) and specificity (true negative rate) were calculated.

[0034] The outcome of the algorithm is the probability that the child shows signs characteristic of autism, defined within range of 0-100%. Result above 50% indicates that the child exhibits early symptoms of autism.

REFERENCES

[0035] Adolph, K. E., Tamis-Lemonda, C. S., Karasik, L. B., 2010. Cinderella indeeda commentary on Iverson's Developing language in a developing body: the relationship between motor development and language development. Journal of Child Language 37, 269-273. doi:10.1017/5030500090999047X [0036] Anzulewicz, A., Sobota, K., Delafield-Butt, J. T., 2016. Toward the autism motor signature: Gesture patterns during smart tablet gameplay identify children with autism. Scientific Reports, 6. doi:10.1038/srep31107 [0037] de Bildt, A., Sytema, S., van Lang, N. D. J., Minderaa, R. B., van Engeland, H., de Jonge, M. V., 2009. Evaluation of the ADOS Revised Algorithm: The Applicability in 558 Dutch Children and Adolescents. J Autism Dev Disord 39, 1350-1358. doi:10.1007/s10803-009-0749-9 [0038] Fabbri-Destro, M., Cattaneo, L., Boria, S., Rizzolatti, G., 2009. Planning actions in autism. Exp Brain Res 192, 521-525. doi:10.1007/s00221-008-1578-3

[0039] Ghaziuddin, M., Butler, E., 1998. Clumsiness in autism and Asperger syndrome: a further report. J Intellect Disabil Res 42 (Pt 1), 43-48. [0040] Kanner, L., 1968. Autistic disturbances of affective contact. Acta Paedopsychiatr 35, 100-136. [0041] Mari, M., Castlello, U., Marks, D., Marraffa, C., Prior, M., 2003. The reach-to-grasp movement in children with autism spectrum disorder. Philosophical Transactions of the Royal Society of London B: Biological Sciences 358, 393-403. doi:10.1098/rstb.2002.1205 [0042] Matson, I L, Mahan, S., Kozlowski, A. M., Shoemaker, M., 2010. Developmental milestones in toddlers with autistic disorder, pervasive developmental disordernot otherwise specified and atypical development. Dev Neurorehabil 13, 239-247. doi:10.3109/17518423.2010.481299 [0043] Ming, X., Brimacombe, M., Wagner, G. C., 2007. Prevalence of motor impairment in autism spectrum disorders, Brain and Development 29, 565-570. doi:10.1016/j.braindev.2007.03.002 [0044] Schmitz, C., Martineau, J., Barthlmy, C., Assaiante, C., 2003. Motor control and children with autism: deficit of anticipatory function? Neuroscience Letters 348, 17-20. doi:10.1016/S0304-3940(03)00644-X [0045] Teitelbaum, P., Teitelbaum, O., Nye, 3., Fryman, I, Maurer, R. G., 1998. Movement analysis in infancy may be useful for early diagnosis of autism. PNAS 95, 13982-13987. doi:10.1.073/pnas.95.23.13982 [0046] Trevarthen, C., 1986. Development of intersubjective motor control in infants., in: Motor Development in Children: Aspects of Coordination and Control. Martinus Nijhoff, Dordrecht, pp. 209-261. [0047] Trevarthen, C., 1984. How Control of Movement Develops, in: Human Motor Actions: Bernstein Reassessed. Elsevier, Amsterdam, pp. 223-261. [0048] Trevarthen, C., Delafield-Butt, J. T., 2013. Autism as a developmental disorder in intentional movement and affective engagement. Front. Integr. Neurosci. 7. doi:10.3389/fnint.2013.00049