Computational methods for the detection of facial palsy

We are developing a telemedicine application which offers automated diagnosis of facial (Bell's) palsy through a Web service. We used a test data set of 43 images of facial palsy patients and 44 normal people to develop the automatic recognition algorithm. Three different image pre-processing methods were used. Machine learning techniques (support vector machine, SVM) were used to examine the difference between the two halves of the face. If there was a sufficient difference, then the SVM recognized facial palsy. Otherwise, if the halves were roughly symmetrical, the SVM classified the image as normal. It was found that the facial palsy images had a greater Hamming Distance than the normal images, indicating greater asymmetry. The median distance in the normal group was 331 (interquartile range 277-435) and the median distance in the facial palsy group was 509 (interquartile range 334-703). This difference was significant (P

A bibliographic snapshot of the telemedicine citation literature

There appear to have been no previous literature-based or literature-oriented studies in telemedicine which have analysed raw citation data. Using a simple search strategy, the Web of Science was analysed up to the end of 2005 to give a snapshot of the field, and to identify matters which would need to be considered in larger scale bibliometric studies. Of the 3673 telemedicine documents retrieved, 2213 (60%) had been cited. Of 56,875 citation records, 32,460 unique citation formats were found. The most-cited paper, and the paper with the greatest annual citation rate, was Perednia and Allen's review article in JAMA, 1995. The two specialist telemedicine journals published 40% of all papers retrieved. In the general literature (i.e. excluding the two specialist journals) there were 1556 citations to their 1374 'citable' articles, apportioned in the ratio 76:24, almost exactly in accordance with the distribution of the articles themselves. However, each of the two specialist telemedicine journals cited itself in a proportion higher than its share of original articles, with an 'excess' of self-citations of 14% in the journal of Telemedicine and Telecare, and 19% in the Telemedicine journal and E-Health. Despite certain technical difficulties, there is considerable scope for bibliometric research in telemedicine.

Successes and failures in assessing acquired language disorders using an Internet-based telerehabilitation system

We have used a telerehabilitation system (eREHAB) to remotely assess acquired language disorders via the Internet. The system was used to establish a 128 kbit/s videoconference between two sites and allowed a remote language assessment to be conducted using the standardized Boston Diagnostic Aphasia Examination (BDAE). The system had the capacity to display text and images, and could play pre-recorded instructions to the participant via various built-in tools. A touch screen allowed tasks involving picture identification to be completed easily. Eighteen participants with a diagnosis of an acquired language disorder were simultaneously assessed using the eREHAB system, and in the traditional face-to-face manner by two speech pathologists. There was very high agreement between the two assessors, with weighted kappa scores of 0.8–1.0 for 88% of the sub-tests of the BDAE. There was also high agreement (80–100%) and high kappa scores (0.67–0.90) between assessors on the six rating scales relating to language characteristics. The agreement between the two assessors for the diagnosis of the type of aphasia was 83%. Limitations of the system related mainly to problems inherent in IP videoconferencing. The inability to maintain the preferred speed of 128 kbit/s for the duration of the videoconference and the resultant increase in video and audio breakup and latency affected the clinician’s ability to administer the BDAE with the same ease and accuracy as in face-to-face administration. These difficulties were exacerbated when participants presented with a moderate to severe language disorder, auditory comprehension deficits or significant hearing loss. Despite these limitations, a valid assessment of language disorder was found to be feasible via this telerehabilitation application.