Cartesian genetic programming for diagnosis of Parkinson disease through handwriting analysis: Performance vs. interpretability issues.

In the last decades, early disease identification through non-invasive and automatic methodologies has gathered increasing interest from the scientific community. Among others, Parkinson's disease (PD) has received special attention in that it is a severe and progressive neuro-degenerative disease. As a consequence, early diagnosis would provide more effective and prompt care strategies, that cloud successfully influence patients' life expectancy. However, the most performing systems implement the so called black-box approach, which do not provide explicit rules to reach a decision. This lack of interpretability, has hampered the acceptance of those systems by clinicians and their deployment on the field. In this context, we perform a thorough comparison of different machine learning (ML) techniques, whose classification results are characterized by different levels of interpretability. Such techniques were applied for automatically identify PD patients through the analysis of handwriting and drawing samples. Results analysis shows that white-box approaches, such as Cartesian Genetic Programming and Decision Tree, allow to reach a twofold goal: support the diagnosis of PD and obtain explicit classification models, on which only a subset of features (related to specific tasks) were identified and exploited for classification. Obtained classification models provide important insights for the design of non-invasive, inexpensive and easy to administer diagnostic protocols. Comparison of different ML approaches (in terms of both accuracy and interpretability) has been performed on the features extracted from the handwriting and drawing samples included in the publicly available PaHaW and NewHandPD datasets. The experimental findings show that the Cartesian Genetic Programming outperforms the white-box methods in accuracy and the black-box ones in interpretability.

Tight data-robust bounds to mutual information combining shuffling and model selection techniques.

The estimation of the information carried by spike times is crucial for a quantitative understanding of brain function, but it is difficult because of an upward bias due to limited experimental sampling. We present new progress, based on two basic insights, on reducing the bias problem. First, we show that by means of a careful application of data-shuffling techniques, it is possible to cancel almost entirely the bias of the noise entropy, the most biased part of information. This procedure provides a new information estimator that is much less biased than the standard direct one and has similar variance. Second, we use a nonparametric test to determine whether all the information encoded by the spike train can be decoded assuming a low-dimensional response model. If this is the case, the complexity of response space can be fully captured by a small number of easily sampled parameters. Combining these two different procedures, we obtain a new class of precise estimators of information quantities, which can provide data-robust upper and lower bounds to the mutual information. These bounds are tight even when the number of trials per stimulus available is one order of magnitude smaller than the number of possible responses. The effectiveness and the usefulness of the methods are tested through applications to simulated data and recordings from somatosensory cortex. This application shows that even in the presence of strong correlations, our methods constrain precisely the amount of information encoded by real spike trains recorded in vivo.

The minimum effective dose of 0.5% hyperbaric spinal bupivacaine for cesarean section.

BACKGROUND: The aim of this prospective, blind study was to determine the minimum effective dose of hyperbaric bupivacaine required for cesarean section. METHODS: With Ethical Committee approval and written consent, 24 healthy women undergoing elective cesarean section received a combined spinal epidural anesthesia. We sought to determine the minimum effective dose of spinal bupivacaine using a staircase method. In each patient an arbitrary dose of 0.5% hyperbaric bupivacaine in relation to patient height was used. The initial dose was 0.075 mg/cm height, while the outcome of each patient's response determined the dose for the subsequent patient. When successful spinal block (sensory level = or < T4 with complete motor blockade) was achieved within 20 min from spinal injection, the dose of spinal bupivacaine for the next patient was decreased by 0.01 mg/cm height. Conversely, when successful spinal block was not observed, the dose of spinal bupivacaine for the next patient was increased by 0.01 mg/cm height. Sensory and motor blocks were evaluated every 5 min by an independent, blinded observer. If successful spinal block was not achieved within the designed period, a 5-8 ml epidural bolus of 2% lidocaine was given to achieve adequate surgical anesthesia. RESULTS: No complications were reported during the study, and all women delivered their baby uneventfully (APGAR scores 5 min after delivery ranged from 9 to 10) within 5 min from uterus incision. The duration of surgical procedure ranged from 30 to 48 minutes. The dose of hyperbaric bupivacaine providing adequate surgical anesthesia within 20 min from spinal injection in 50% of subjects was 0.036 mg/cm height (95% confidence intervals: 0.031-0.041 mg/cm height). The ED95 calculated from the probit transformation to provide effective spinal anesthesia for cesarean section was 0.06 mg/cm height. CONCLUSIONS: This prospective, blind study demonstrated that a dose as low as 0.06 mg/cm height represents the dose of intrathecal bupivacaine providing effective spinal block in 95% of women undergoing elective cesarean section.

Effects of speed of intrathecal injection on unilateral spinal block by 1% hyperbaric bupivacaine. A randomized, double-blind study.

BACKGROUND: To evaluate if the speed of intrathecal injection affects the lateral distribution of spinal block during unilateral spinal anaesthesia with 1% hyperbaric bupivacaine. DESIGN: prospective, randomized, double-blind study. SETTING: anaesthesia Department at a University Hospital. PATIENTS: 30 ASA physical status I-II patients, scheduled for elective orthopedic surgery involving one lower limb only (ankle and foot surgery). INTERVENTIONS: after placing the patients in the lateral decubitus position with the site to be operated on dependent, dural puncture was performed at L3-L4 interspace using a 25-Gauge Whitacre spinal needle. After the needle hole had been turned toward the dependent side, patients were randomized to receive 8 mg of 1% hyperbaric bupivacaine injected over either 40 sec (Group SLOW, n = 15) or 3 sec (Group FAST, n = 15). The lateral position was maintained for 15 min. MEASUREMENTS: a blind observer evaluated the evolution of sensory (pinprick test) and motor (modified Bromage scale) blocks on both the dependent and nondependent sides until the regression of motor block by one degree. RESULTS: No differences in the maximal sensory level on both the dependent and nondependent sides were observed in the two groups. Unilateral sensory block was onserved in 6 patients in group SLOW (40%) and in 5 patients in group FAST (33%), (p = not significant[NS]. Unilateral motor block was observed in 9 patients in group SLOW (60%) and in 10 patients in group FAST (66%) (NS). CONCLUSIONS: Extremely reduced speeds of intrathecal injection did not improve the lateral distribution of spinal block when injecting 8 mg of 1% hyperbaric bupivacaine through a pencil-point directional spinal needle.

Anaesthesia for ultrasound guided oocyte retrieval: midazolam/remifentanil versus propofol/fentanyl regimens.

To evaluate the quality of intra-operative anaesthesia and recovery characteristics of two different anaesthesia regimens, 60 healthy women undergoing ultrasound guided oocyte retrieval for in vitro fertilization procedures were randomly allocated to receive either a propofol/fentanyl or a midazolam/remifentanil based anaesthesia. The surgical procedure was successful in all patients and no severe side effects were reported by any patient. Four patients in the midazolam/remifentanil group (13%) would not accept the same anaesthetic procedure for further in vitro fertilization treatment due to intra-operative awareness, while all propofol/fentanyl patients were prepared to accept the same procedure again (P < 0.05). Patients in the propofol/fentanyl group required manual ventilation more frequently through a facemask than those patients treated with the midazolam/remifentanil combination (50% and 30%, respectively; P < 0.05). The time to achieve an Aldrete's score of 10 was shorter in the midazolam/remifentanil patients (2 +/- 2 min) than in those who received propofol/fentanyl (4 +/- 2 min) (P < 0.001), but no differences were observed in the time required to be 'fit to discharge' from the post-anaesthesia care unit. We conclude that the use of a midazolam/remifentanil regimen is as effective and safe as a fentanyl/propofol regimen in patients undergoing transvaginal oocyte retrieval for in vitro fertilization procedures.

Block distribution and cardiovascular effects of unilateral spinal anaesthesia by 0.5% hyperbaric bupivacaine. A clinical comparison with bilateral spinal block.

BACKGROUND: A low dose of hyperbaric local anaesthetic solution, pencil point needle and slow speed of intrathecal injection have been reported to obtain a unilateral distribution of spinal anaesthesia. This should also minimize cardiovascular effects of spinal block. The aim of this prospective, randomized, parallel group study was to evaluate cardiac performance during unilateral subarachnoid block and to compare it with that produced by standard bilateral spinal anaesthesia. METHODS: With Ethic Committee approval and patients' consent we studied 30 ASA I-II patients scheduled for one leg surgery. Dural puncture was performed by 25-G Whitacre needle with patients lying in the lateral position and the side to be operated on dependent. Patients then randomly received 8 mg of 0.5% hyperbaric bupivacaine injected over 80 sec with needle hole orientated towards the dependent side (Unilateral, n = 15), or 15 mg of the same solution injected over 6 sec with needle bevel cranially directed (Control, n = 15). Only patients of the Unilateral group remained in the lateral position for 15 min. Noninvasive Arterial blood pressure, heart rate, stroke volume index and cardiac index were measured before spinal block (baseline) and then at 5, 15, 30 and 45 min; while sensory and motor blocks were evaluated at 15, 30 and 45 min on both sides. RESULTS: Unilateral spinal anaesthesia was observed in 11 patients of the Unilateral group (73%). Three patients of the Control group (20%) required colloids for intraoperative hypotension. Mean arterial blood pressure and heart rate decreased from baseline only in the Control group (p = 0.001 and p = 0.0003 respectively), while heart rate was decreased in Control even when compared to Unilateral group (p = 0.01). The stroke volume index was unchanged in the two groups (p = 0.22), while the cardiac index showed a 15-20% decrease from baseline in Control group (p = 0.001), with a significant decrease at 30 and 45 min when compared to Unilateral (p = 0.01). DISCUSSION: The use of 8 mg of 0.5% hyperbaric bupivacaine slowly injected through a directional needle provided a spinal block relatively restricted to the operative side with minimal effects on cardiovascular homeostasis.

Cardiorespiratory changes during gynaecological laparoscopy by abdominal wall elevation: comparison with carbon dioxide pneumoperitoneum.

We have studied the cardiorespiratory changes produced by abdominal wall elevation (AWE) or carbon dioxide pneumoperitoneum (PN) in 20 women undergoing gynaecological laparoscopy. Arterial pressure, heart rate, lung/chest complicance and blood-gas tensions were measured 10 min after induction of general anaesthesia (T0), 10 min after abdominal distension in the supine position (T1) and 10 min after the Trendelenburg position was assumed (T2). Visual analogue scores for pain were recorded 1 and 6 h after the end of surgery. We found that lung/chest compliance was reduced significantly in group PN at T1 and T2 compared with both T0 and group AWE. Diastolic arterial pressure increased significantly in group PN at T1 and T2 compared with both T0 and group AWE, while it remained unchanged in group AWE. Arterial PCO2 increased significantly only in group PN after pneumoperitoneum, while oxygenation was almost unchanged in both groups. AWE patients had greater abdominal pain 1 h after surgery. Six hours after surgery pain was similar in the two groups. These data indicate that abdominal wall elevation reduced pulmonary compliance less than a pneumoperitoneum in patients undergoing gynaecological laparoscopy.

Dialysis treatment using an ethylene vinyl alcohol membrane and no anticoagulation for chronic uremic patients.

Anticoagulation used in hemodialysis treatment brings with it the risk of hemorrhagic complications and the possible consequences associated with chronic heparin administration. These problems have not been satisfactorily addressed to date. This study examined a new dialysis method that does not require the administration of anticoagulants. Dialysis is performed for 3 h with a continuous infusion of 750 ml of physiological saline solution in predilution mode and using filters made of ethylene vinyl alcohol copolymer membranes. Eleven patients with chronic uremia underwent more than 2,000 dialysis treatments performed with 9 episodes of coagulation of the dialyzer or blood tubings (0.43%). An evaluation of individual treatments revealed a high degree of biocompatibility and only a scanty activation of coagulation. Blood depuration efficacy was very good as evaluated from pretreatment and posttreatment routine blood chemistries. The technique described here represents a simple and effective method for performing regular dialysis treatment and does not require anticoagulant therapy.

Friedreich's ataxia: clinical involvement and evoked potentials.

In 15 patients definitely affected by Friedreich's ataxia, precocious SEP abnormalities were constantly recorded independent of the duration of symptoms and the severity of clinical involvement; VEPs were frequently involved, but neither VEP abnormalities, nor visual impairment could be correlated with the severity or duration of Friedreich's ataxia clinical involvement; BAEPs were, to different extents, abnormal. These were completely dissociated from hearing disorders and significantly correlated (P less than 0.001) with Friedreich's ataxia clinical disability. Our observation of increased latency of the VEP P100 component and the early disappearance of BAEP Wave V with persistence of Wave I is in contrast with the hypothesis that changes are related to primary axonal degeneration. For practical purposes SEPs could help in the early diagnosis of Friedreich's ataxia, and BAEPs could be used for the electrophysiological monitoring of its clinical progression.

Stability of peritoneal urea clearance in continuous ambulatory peritoneal dialysis (18 months experience with CAPD).

A group of 13 patients on Continous Ambulatory Peritoneal Dialysis up to 18 months was studied. During the study period, Peritoneal Urea Clearance was stable. In all but 2 cases peritonitis did not depress the efficiency in clearing urea. When Peritoneal Urea Clearance was corrected for 1.73 m2 surface area, there was on the average an 11.8% increase over not corrected clearance rates. This increase was statistically significant.

Effects of catabolic stress in acute and chronic renal failure.

The degree of catabolism was studied in a group of seven patients with postsurgical acute renal failure, and net protein catabolism of 43.1 to 927 g/day. In 70% of a group of 30 patients with stable chronic renal failure receiving and 18 g high biological value protein diet, urea synthesis was equivalent to what could be accounted for by the protein content of the food. This equivalence was not found in studies in both adults and children who were treated for catabolic chronic renal failure. Data obtained show the superiority of feeding essential amino acids with an adequated caloric supply. The need for studies aimed at adapting the uremic patient to protein depletive mechanisms is strongly stressed.