PubMed Health⌕ Search

Biomedical subjects

S M Lo

Publications and source records attributed to S M Lo.

6 recordsLinked to original sources
PubMed2004

A hybrid neural network model for noisy data regression.

A hybrid neural network model, based on the fusion of fuzzy adaptive resonance theory (FA ART) and the general regression neural network (GRNN), is proposed in this paper. Both FA and the GRNN are incremental learning systems and are very fast in network training. The proposed hybrid model, denoted as GRNNFA, is able to retain these advantages and, at the same time, to reduce the computational requirements in calculating and storing information of the kernels. A clustering version of the GRNN is designed with data compression by FA for noise removal. An adaptive gradient-based kernel width optimization algorithm has also been devised. Convergence of the gradient descent algorithm can be accelerated by the geometric incremental growth of the updating factor. A series of experiments with four benchmark datasets have been conducted to assess and compare effectiveness of GRNNFA with other approaches. The GRNNFA model is also employed in a novel application task for predicting the evacuation time of patrons at typical karaoke centers in Hong Kong in the event of fire. The results positively demonstrate the applicability of GRNNFA in noisy data regression problems.

Journal Article↗
PubMed2001

Interstitial adenosine concentration in rat red or white skeletal muscle during systemic hypoxia or contractions.

Interstitial adenosine concentrations in red soleus (SL) or white extensor digitorum longus (EDL) muscles of anaesthetised rats were determined using microdialysis and HPLC. Systemic hypoxia was induced by ventilating the animals with 10% oxygen in nitrogen for 15 min: arterial PO2 decreased from 111.8 +/- 10.9 to 42.2 +/- 4.3 mmHg (n = 4; P < 0.01) and mean systemic arterial blood pressure from 97.6 +/- 4.9 to 59.0 +/- 3.6 mmHg (n = 22; P < 0.001). The interstitial adenosine concentration was not significantly changed from its control values of 294 +/- 44 nM (n = 20) in EDL and 302 +/- 36 nM (n = 20) in SL during hypoxia or the recovery period. The interstitial lactate concentration did not change in the early part of the hypoxia but increased from 1.0 +/- 0.2 to 1.4 +/- 0.3 mM (n = 6; P < 0.05) in SL and from 2.0 +/- 0.4 to 2.4 +/- 0.4 mM (n = 6; P < 0.05) in EDL during the later part of the hypoxia, and remained elevated in the recovery period. Muscle contractions (2 Hz for 15 min) produced a transient increase in the interstitial adenosine concentration of SL from 150 +/- 35 to 244 +/- 75 nM (n = 10; P < 0.05) during the first 5 min of stimulation. In EDL the interstitial adenosine concentration increased from 145 +/- 50 to 435 +/- 144 nM (n = 10; P < 0.05) in the later part of the contraction and remained elevated in the early part of the recovery period. These data suggest that: (i) in systemic hypoxia adenosine does not appear in the interstitial space, which rules out its release from skeletal muscle, although it may be formed by the vascular tissues in this condition; (ii) adenosine is formed in the interstitial space of skeletal muscle during muscle contractions; (iii) there is slow clearance of adenosine from the interstitial space of white muscle, perhaps due to the low vascularity of the tissue.

Adenosine↗
PubMed1998

Tumour volume: implications in T2/T3 glottic/supraglottic squamous cell carcinoma.

OBJECTIVE: In this study, tumour volume was investigated to determine if it predicts locoregional control of T2/T3 glottic/supraglottic laryngeal carcinoma treated with radiotherapy or surgery. The effect of radiotherapy dosage was also assessed in those patients treated with primary radiotherapy. The ability to identify a subset of patients suitable for primary radiotherapy and, hence, voice preservation based on pretreatment computerized axial tomography (CT) would be valuable. METHOD: The charts of 55 patients referred to the London Regional Cancer Centre (LRCC) between 1988 to 1994 were reviewed. Each presented with a previously untreated T2 or T3 squamous cell carcinoma (SCC) of either the glottic or supraglottic larynx. Tumour volume was calculated from pretreatment CT scans by observers unaware of the clinical data associated with each radiograph. Wilcoxon t test and univariate and multivariate Cox regression analyses were performed. RESULTS: Mean tumour volume differed between those patients treated with radiotherapy and those treated surgically (4.5 cm3 and 11 cm3, respectively; p < .01). Mean tumour volume also differed between T2 and T3 tumours in the primary radiotherapy group (3.8 cm3 and 9.3 cm3, respectively; p < .01). Tumour volume > 4.0 cm3 was a significant predicator of local failure in T2 laryngeal tumours treated with radiotherapy (p < .05). This volume effect was not abolished with increasing radiotherapy dosage. Tumour volume was not a significant predictor of local control in the T3 tumours treated with radiotherapy or when all tumours, irrespective of T stage, that were treated with radiotherapy were considered together. There was no similar volume effect found in the surgical group. CONCLUSIONS: Tumour volume > 4 cm3 predicts local failure in T2 laryngeal tumours treated with radiotherapy regardless of radiotherapy dosage. This volume effect is not seen in those tumours treated with surgery. Inclusion of tumour volume data may eventually augment our current classification system.

Adult↗