A network analysis of developing brain cultures.

We recorded electrical activity from four developing embryonic brain cultures (4-40 days in vitro) using multielectrode arrays (MEAs) with 60 embedded electrodes. Data were filtered for local field potentials (LFPs) and downsampled to 1 ms to yield a matrix of time series consisting of 60 electrode × 60 000 time samples per electrode per day per MEA. Each electrode time series was rendered stationary and nonautocorrelated by applying an ARIMA (25, 1, 1) model and taking the residuals (i.e. innovations). Two kinds of analyses were then performed. First, a pairwise crosscorrelation (CC) analysis (±25 1 ms lags) revealed systematic changes in CC with lag, day in vitro (DIV), and inter-electrode distance. Specifically, (i) positive CCs were 1.76× more prevalent and 1.44× stronger (absolute value) than negative ones, and (ii) the strength of CC increased with DIV and decreased with lag and inter-electrode distance. Second, a network equilibrium analysis was based on the instantaneous (1 ms resolution) logratio of the number of electrodes that were above or below their mean, called simultaneous departure from equilibrium, SDE. This measure possesses a major computational advantage over the pairwise crosscorrelation approach because it is very simple and fast to calculate, an important factor for the analysis of large networks. The results obtained with SDE covaried highly with CC over DIV, which further validates the usefulness of this measure as a computationally effective tool for large scale network analysis.

Daptomycin, fosfomycin, or both for treatment of methicillin-resistant Staphylococcus aureus osteomyelitis in an experimental rat model.

The in vivo activities of daptomycin, fosfomycin, and a combination of both antibiotics against a clinical isolate of methicillin-resistant Staphylococcus aureus (daptomycin MIC, 0.25 µg/ml; fosfomycin MIC, 0.5 µg/ml) were evaluated in a rat model of osteomyelitis. A total of 37 rats with experimental osteomyelitis were treated for 4 weeks with either 60 mg/kg of body weight of daptomycin subcutaneously once daily, 75 mg/kg fosfomycin intraperitoneally once daily, a combination of both drugs, or a saline placebo. After the completion of treatment, animals were euthanized, and the infected tibiae were processed for quantitative bacterial culture. Bone cultures were found to be positive for methicillin-resistant S. aureus in 9 of 9 (100%) animals of the placebo group, in 9 of 9 (100%) animals treated with daptomycin, in 1 of 10 (10%) fosfomycin-treated rats, and in 1 of 9 (22.2%) rats comprising the combination group. Results of bacterial counts in the bone samples were expressed as log(10) CFU/g of bone and analyzed by using the Mann-Whitney U test followed by Bonferroni's multiple-comparison test. Based on bacterial counts, treatment with daptomycin was significantly superior to placebo, although it remained inferior to treatment with fosfomycin. No synergistic or antagonistic effect was observed for the combination therapy. No development of resistance against daptomycin or fosfomycin was observed after the 4-week treatment period.

True associations between resting fMRI time series based on innovations.

We calculated voxel-by-voxel pairwise crosscorrelations between prewhitened resting-state BOLD fMRI time series recorded from 60 cortical areas (30 per hemisphere) in 18 human subjects (nine women and nine men). Altogether, more than a billion-and-a-quarter pairs of BOLD time series were analyzed. For each pair, a crosscorrelogram was computed by calculating 21 crosscorrelations, namely at zero lag ± 10 lags of 2 s duration each. For each crosscorrelogram, in turn, the crosscorrelation with the highest absolute value was found and its sign, value, and lag were retained for further analysis. In addition, the crosscorrelations at zero lag (irrespective of the location of the peak) were also analyzed as a special case. Based on known varying density of anatomical connectivity, we distinguished four general brain groups for which we derived summary statistics of crosscorrelations between voxels within an area (group I), between voxels of paired homotopic areas across the two hemispheres (group II), between voxels of an area and all other voxels in the same (ipsilateral) hemisphere (group III), and voxels of an area and all voxels in the opposite (contralateral) hemisphere (except those in the homotopic area) (group IV). We found the following. (a) Most of the crosscorrelogram peaks occurred at zero lag, followed by ± 1 lag; (b) over all groups, positive crosscorrelations were much more frequent than negative ones; (c) average crosscorrelation was highest for group I, and decreased progressively for groups II-IV; (d) the ratio of positive over negative crosscorrelations was highest for group I and progressively smaller for groups II-IV; (e) the highest proportion of positive crosscorrelations (with respect to all positive ones) was observed at zero lag; and (f) the highest proportion of negative crosscorrelations (with respect to all negative ones) was observed at lag = 2. These findings reveal a systematic pattern of crosscorrelations with respect to their sign, magnitude, lag and brain group, as defined above. Given that these groups were defined along a qualitative gradient of known overall anatomical connectivity, our results suggest that functional interactions between two voxels may simply reflect the density of such anatomical connectivity between the areas to which the voxels belong.

Efficacy of fosfomycin in experimental osteomyelitis due to methicillin-resistant Staphylococcus aureus.

The activity of fosfomycin was evaluated in an experimental methicillin-resistant Staphylococcus aureus (MRSA) osteomyelitis model. Eighteen rats were treated for 4 weeks with 150 mg of fosfomycin/kg of body weight intraperitoneally once daily or with saline placebo. After treatment, animals were euthanized and the infected tibiae were processed for quantitative bacterial culture. Bone cultures were positive for methicillin-resistant S. aureus in all 9 (100%) untreated controls and in 2 of 9 (22.2%) fosfomycin-treated rats. Thus, fosfomycin treatment was significantly more efficacious than placebo. No development of resistance was observed after the 4-week treatment period.

Post-traumatic stress disorder: a right temporal lobe syndrome?

In a recent paper (Georgopoulos et al 2010 J. Neural Eng. 7 016011) we reported on the power of the magnetoencephalography (MEG)-based synchronous neural interactions (SNI) test to differentiate post-traumatic stress disorder (PTSD) subjects from healthy control subjects and to classify them with a high degree of accuracy. Here we show that the main differences in cortical communication circuitry between these two groups lie in the miscommunication of temporal and parietal and/or parieto-occipital right hemispheric areas with other brain areas. This lateralized temporal-posterior pattern of miscommunication was very similar but was attenuated in patients with PTSD in remission. These findings are consistent with observations (Penfield 1958 Proc. Natl Acad. Sci. USA 44 51-66, Penfield and Perot 1963 Brain 86 595-696, Gloor 1990 Brain 113 1673-94, Banceaud et al 1994 Brain 117 71-90, Fried 1997 J. Neuropsychiatry Clin. Neurosci. 9 420-8) that electrical stimulation of the temporal cortex in awake human subjects, mostly in the right hemisphere, can elicit the re-enactment and re-living of past experiences. Based on these facts, we attribute our findings to the re-experiencing component of PTSD and hypothesize that it reflects an involuntarily persistent activation of interacting neural networks involved in experiential consolidation.

The synchronous neural interactions test as a functional neuromarker for post-traumatic stress disorder (PTSD): a robust classification method based on the bootstrap.

Traumatic experiences can produce post-traumatic stress disorder (PTSD) which is a debilitating condition and for which no biomarker currently exists (Institute of Medicine (US) 2006 Posttraumatic Stress Disorder: Diagnosis and Assessment (Washington, DC: National Academies)). Here we show that the synchronous neural interactions (SNI) test which assesses the functional interactions among neural populations derived from magnetoencephalographic (MEG) recordings (Georgopoulos A P et al 2007 J. Neural Eng. 4 349-55) can successfully differentiate PTSD patients from healthy control subjects. Externally cross-validated, bootstrap-based analyses yielded >90% overall accuracy of classification. In addition, all but one of 18 patients who were not receiving medications for their disease were correctly classified. Altogether, these findings document robust differences in brain function between the PTSD and control groups that can be used for differential diagnosis and which possess the potential for assessing and monitoring disease progression and effects of therapy.

National paediatric immunization program of high risk groups: no effect on the incidence of invasive pneumococcal diseases.

This study monitors the epidemiology of invasive pneumococcal diseases (IPD) in hospitalized children up to 60 months of age before (February 2001-October 2004) and after (November 2004-January 2007) the introduction of a national risk group immunization program with "Prevenar" in Austria. The IPD incidence rates, per 100,000, for IPD were 7.6 before and 6.4 after the risk group immunization program, while there was a significant reduction (p<0.05) for meningitis, 3.1 before and 1.6 after. Overall, the most commonly observed serotypes were 14 (34.2%), 6B (11.7%), and 23F (6.7%). 71.7% of the identified strains were vaccine types; 12.5% were vaccine-related serotypes. No clinically relevant changes in the incidence rate of IPDs or shift/replacement of serotypes was documented. Antimicrobial resistance predominated against erythromycin (32.5%) and clarithromycin (26.7%). Our data show that this risk group vaccination program had no impact on the incidence of IPD in young children.

The impact of an erbium, chromium:yttrium-scandium-gallium-garnet laser with radial-firing tips on endodontic treatment.

Radial-firing tips should allow a more homogeneous laser irradiation of root canal walls. The aim of the study was to assess the effects of erbium, chromium:yttrium-scandium-gallium-garnet (Er,Cr:YSGG) laser irradiation in conjunction with those newly designed tips. The investigation comprised bacteriology, morphological evaluations and temperature measurements. Root canals were inoculated with two test strains and laser irradiated with power settings of 0.6 W and 0.9 W and a repetition rate of 20 Hz. Subsequently, the samples were subjected to microbiological evaluation. The morphological changes of the canal walls were assessed by scanning electron microscopy. To reveal possible thermal side effects, we carried out temperature measurements. The bacteriological evaluation revealed a decisive disinfectant effect. Scanning electron microscopy showed the homogeneous removal of smear layer from the root canal walls. The temperature rise at the root surface during the irradiation was moderate, yielding 1.3 degrees C for the 0.6 W setting and 1.6 degrees C for the 0.9 W setting. The investigations indicated that the Er,Cr:YSGG laser, in conjunction with radial-firing tips, is a suitable tool for the elimination of bacteria in root canals and for the removal of smear layer.

In vitro antimicrobial activity of the novel polymeric guanidine Akacid plus.

The bactericidal and fungicidal activity of Akacid plus, a novel polymeric compound of the cationic family of disinfectants, was compared with chlorhexidine digluconate using quality control strains of Staphylococcus aureus, Enterococcus hirae, Escherichia coli, Pseudomonas aeruginosa, Candida albicans and Aspergillus niger. In vitro activity was determined using the quantitative suspension tests described by the European Committee for Standardization. These use concentrations of 0.01-0.5% against bacterial strains/C. albicans, with 0.5-4% against A. niger, and exposure times of 5, 15 and 60 min in the presence and absence of 0.3% bovine albumin and with dilution in distilled and hard water. In the basic quantitative suspension test, Akacid plus destroyed all bacterial pathogens at a concentration of 0.1% in < or =5 min. Chlorhexidine was also highly active against S. aureus, E. coli and P. aeruginosa, but failed to eliminate E. hirae within 5 min. A high organic load reduced the bactericidal activity of both disinfectants slightly. Akacid plus showed fungicidal activity against C. albicans within 15-60 min and eliminated A. niger at a concentration of 1% in 5 min of contact. Chlorhexidine was fungicidal against C. albicans, but not against A. niger.

Plasma concentrations might lead to overestimation of target site activity of piperacillin in patients with sepsis.

OBJECTIVES: Pharmacokinetic (PK)/pharmacodynamic (PD) models have become increasingly important in optimizing antimicrobial therapy. This approach is highly recommended by regulatory authorities intending to force the evaluation of antimicrobial action at the site of infection. METHODS: Clinical isolates of Pseudomonas aeruginosa and Staphylococcus aureus with MICs of 4, 8 and 16 mg/L for piperacillin were used in an in vivo PK/in vitro PD model. Bacteria were exposed in vitro to the concentration-versus-time profiles of piperacillin in plasma and subcutaneous adipose tissue measured in vivo in septic patients. Samples were withdrawn at defined intervals and the numbers of bacteria per mL were counted and plotted against time. RESULTS: Piperacillin levels determined in plasma were able to effectively inhibit bacterial growth of all bacterial strains used in the present study (MIC ranged from 4-16 mg/L). In contrast, concentration-versus-time profiles of subcutaneous adipose tissue were effective in killing isolates with MICs of 4 and 8 mg/L only, while bacterial growth of S. aureus and P. aeruginosa with MICs of 16 mg/L was not inhibited. CONCLUSIONS: Bacteria with MICs < 16 mg/L were effectively inhibited in subcutaneous adipose tissue in patients with sepsis. The prediction of microbiological outcome based on concentrations of piperacillin in plasma resulted in a marked overestimation of antimicrobial activity at the site of infection.

Development of simple thiol-reactive liposome formulations, one-step analysis and physicochemical characterization.

The aim of this study was to synthesize simple thiol-reactive conjugates from maleimide and lipoamines (stearylamine or oleylamine) and to develop a simple, fast and low-cost method for the preparation of lyophilized general-purpose thiol-reactive liposomes. A formulation of egg phosphatidylcholine-dipalmitoylphoshatidylglycerol (9:0.1 molar ratio) was developed and characterized. Freeze-drying methodology was established to produce a stock of liposomes and the physicochemical characteristics of the reconstituted liposomes were compared with those of the initial preparation. The physicochemical properties (size and zeta potential) of the new liposomal formulations were studied. High-performance thin-layer chromatography coupled to a flame ionization detector was applied for one-step analysis of the liposomal components and for determining the maleimide-lipoamine conjugates phospholipid molar ratio. The differences concerning the incorporation efficiency of the synthetic conjugates into liposomes were discussed on the basis of their conformational properties. The small difference in structure between the two thiol-reactive conjugates (i.e., the C18 alkyl chain double bond) causes a considerable difference in phospholipids packing of the resulting lipidic bilayers of the liposomes; the conformational bending of conjugate maleimide-oleylamine may contribute to the final architecture of liposomes.

A comparative study of the effects of cholesterol and sclareol, a bioactive labdane type diterpene, on phospholipid bilayers.

Sclareol (labd-14-ene-8,13-diol) is a highly water-insoluble molecule that belongs to the labdane type diterpenes and is characterized as a biologically active molecule, due to its cytotoxic and cytostatic effects against human leukemic cell lines. A superimposition study between sclareol and cholesterol, based on their corresponding hydrophobic and polar molecular segments calculated from their lipophilic profiles, revealed their spatial similarities. This structural similarity between the two molecules prompted us to compare their effects on the structure and stability of phospholipid dipalmitoylphosphatidylcholine (DPPC) membranes. Differential scanning calorimetry (DSC) was applied to compare the thermal changes caused by either cholesterol or sclareol when are incorporated in DPPC bilayers. The results showed that sclareol is incorporated into phospholipid model membranes and mimics the thermal effects of cholesterol especially at concentrations up to X(sclareol)=9.1 mol%. These effects can be summarized as the abolition of pre-transition, lowering of the main phase transition and reduction of the enthalpy change (DeltaH) of the gel to liquid-crystalline phase transition of DPPC bilayers. At concentrations X> or =16.7 mol%, sclareol and cholesterol caused different heterogeneity in lipid bilayers or a reversible transition from a vesicular suspension to an extended peak bilayer network. This different fluidization, exerted by the two molecules at high concentration, may be related to their different stability and the z-average mean diameter of the liposomes they form. Small unilamellar vesicles, prepared by the thin film hydration method showed that DPPC bilayers containing a high concentration of sclareol in equimolar ratio sclareol:cholesterol were unstable, in contrast to the ones containing only cholesterol.

Viridans streptococci in endocarditis and neutropenic sepsis: biofilm formation and effects of antibiotics.

OBJECTIVES: Viridans group streptococci (VGS) are a frequent cause of bacterial endocarditis or sepsis in patients with neutropenia. Endocarditis in particular, is associated with plaque formation on the endocardium and valve leaflets whereas VGS septicaemia in neutropenic patients is caused by the influx of oral flora bacteria through mucositic lesions. This study examined the in vitro potency for biofilm formation of clinical VGS bloodstream isolates, and the effects of antibiotics on these biofilms. METHODS: During the years 1998-2000, 40 VGS bloodstream isolates from 18 patients with endocarditis and 22 patients with severe sepsis and neutropenia were collected. The MICs of penicillin, teicoplanin and moxifloxacin were determined using the microdilution broth method according to NCCLS criteria. Biofilms were grown in microtitre plates, dyed with Crystal Violet, and the mean optical density (OD) was used for quantification. Biofilms were incubated with penicillin, teicoplanin and moxifloxacin at various concentrations starting with the MICs for the respective isolates tested. RESULTS: Isolates from eight out of 18 patients with endocarditis and six out of 22 patients with neutropenia formed biofilms (not significant). For the 14 isolates, the MIC(90)s (range) of penicillin, teicoplanin and moxifloxacin were 0.5 mg/L (0.001-0.5), 0.125 mg/L (0.025-0.125) and 0.5 mg/L (0.05-0.5), respectively. Generally, biofilms persisted although incubated with the antibiotics up to concentrations of 128 x MIC. However, the ODs of biofilms after incubation with an antibiotic were significantly lower than the ODs of biofilms without antibiotic (P<0.05). A significant decrease in the biofilms with increasing antibiotic concentrations was observed for teicoplanin and moxifloxacin, but not for penicillin G. CONCLUSIONS: VGS isolated from patients with endocarditis and patients with sepsis and neutropenia form biofilms. Biofilms persist even when exposed to antibiotics at concentrations up to 128 x MIC. Nevertheless, teicoplanin and moxifloxacin reduced the density of the biofilms at concentrations >/=16 x MIC. Thus, testing the effects of antibiotics on biofilms may supply useful information in addition to standard in vitro testing, particularly in diseases where biofilm formation is involved in the pathogenesis.

Impact of plasma protein binding on antimicrobial activity using time-killing curves.

OBJECTIVES: Plasma protein binding (PPB) is known to impair the antimicrobial activity of beta-lactams, but its impact on the activity of other classes of antimicrobials such as fluoroquinolones is controversial. This study was undertaken to investigate the effect of PPB on bacterial killing by selected antibiotics and moxifloxacin, which served as a model compound for the class of fluoroquinolones. METHODS: Bacterial time-killing curves were employed in the absence and presence of physiological albumin concentrations (40 g/L). Moxifloxacin, ampicillin and oxacillin were investigated. Fosfomycin, a non-protein bound antibiotic was used for comparison. Simulations were carried out by employing concentrations of antibiotics of one-fourth of the minimal inhibitory concentration (MIC), equal to the MIC and four-fold the MIC of one select bacterial strain (Staphylococcus aureus ATCC 29213). To correlate bacterial killing to the extent of PPB, bacterial time-killing curves were plotted using the calculated free and the total drug concentration. RESULTS: Bacterial killing by fosfomycin was not affected by the addition of albumin. The antimicrobial activity of oxacillin and ampicillin was reduced in the presence of albumin as expected by the calculation of the free fraction of these antibiotics. Adding albumin to moxifloxacin resulted in a significant decrease in bacterial killing of more than 1 log10 cfu/mL after a period of 8 h when the moxifloxacin concentration was equal to the respective MIC. CONCLUSIONS: Our data confirm the view that albumin substantially impairs the antimicrobial activity of antibiotics including moxifloxacin, a member of the class of fluoroquinolones.

Neurophysiology of the parieto-frontal system during target interception.

We studied the functional properties of neurons of two elements of the parieto-frontal system: area 7a of the PPC and the motor cortex (M1), during an interception task of stimuli moving in real (RM) and apparent motion (AM). The stimulus moved along a circular path with one of 5 speeds, and was intercepted at 6 o'clock by exerting a force pulse on a joystick. A smooth stimulus motion was produced in RM, whereas in AM 5 stimuli were flashed successively at the vertices of a pentagon. The results showed, that a group of neurons in both areas above responded not only during the interception but also during a NOGO task in which the same stimuli were presented in the absence of a motor response. Most of these neurons were tuned to the stimulus angular position. In addition, we found that the time-varying neuronal activity in both areas was related to various aspects of stimulus motion and hand force, with stimulus-related activity prevailing in area 7a and hand-related activity prevailing in M1. Interestingly, the neural activity was selectively associated with the stimulus angle during RM, whereas it was tightly correlated to the time-to-contact during AM. Thus, the results suggest that area 7a was processing high level features of the circularly moving stimuli and was involved in the production of an early command signal for stimulus interception, whereas M1 was still processing some aspect of the visual stimulus that were used to trigger the interception movement using a predictive mechanism.

Relevance of soft-tissue penetration by levofloxacin for target site bacterial killing in patients with sepsis.

Antimicrobial therapy of soft tissue infections in patients with sepsis sometimes lacks efficiency, despite the documented susceptibility of the causative pathogen to the administered antibiotic. In this context, impaired equilibration between the antibiotic concentrations in plasma and those in tissues in critically ill patients has been discussed. To characterize the impact of tissue penetration of anti-infective agents on antimicrobial killing, we used microdialysis to measure the concentration-versus-time profiles of levofloxacin in the interstitial space fluid of skeletal muscle in patients with sepsis. Subsequently, we applied an established dynamic in vivo pharmacokinetic-in vitro pharmacodynamic approach to simulate bacterial killing at the site of infection. The population mean areas under the concentration-time curves (AUCs) for levofloxacin showed that levofloxacin excellently penetrates soft tissues, as indicated by the ratio of the AUC from time zero to 8 h (AUC(0-8)) for muscle tissue (AUC(0-8 muscle)) to the AUC(0-8) for free drug in plasma (AUC(0-8 plasma free)) (AUC(0-8 muscle)/AUC(0-8 plasma free) ratio) of 0.85. The individual values of tissue penetration and maximum concentration (C(max)) in muscle tissue were highly variable. No difference in bacterial killing of a select Staphylococcus aureus strain for which the MIC was 0.5 microg/ml was found between individuals after exposure to dynamically changing concentrations of levofloxacin in plasma and tissue in vitro. In contrast, the decrease in the bacterial counts of Pseudomonas aeruginosa (MIC = 2 microg/ml) varied extensively when the bacteria were exposed to levofloxacin at the concentrations determined from the individual concentration-versus-time profiles obtained in skeletal muscle. The extent of bacterial killing could be predicted by calculating individual C(max)/MIC and AUC(0-8 muscle)/AUC(0-8 plasma free) ratios (R = 0.96 and 0.93, respectively). We have therefore shown in the present study that individual differences in the tissue penetration of levofloxacin may markedly affect target site killing of bacteria for which MICs are close to 2 microg/ml.

Target site bacterial killing of cefpirome and fosfomycin in critically ill patients.

We employed an in-vivo pharmacokinetic/in-vitro pharmacodynamic method to simulate bacterial killing in plasma and the interstitium of skeletal muscle tissue after intravenous administration of 2 g of cefpirome and 8 g of fosfomycin alone and in combination to patients with sepsis. Interstitial antimicrobial concentrations were determined by use of in-vivo microdialysis. CFU/ml of Staphylococcus aureus (ATCC 29213) and Pseudomonas aeruginosa (clinical isolate) decreased by approximately 2log(10) for plasma and muscle tissue 6 h after cefpirome and fosfomycin administration compared with the baseline, respectively. The simulation of plasma and tissue pharmacokinetics for the combined administration of these antibiotics resulted in complete eradication of S. aureus within 5 h after drug exposure. No bacterial re-growth occurred in any of the simulations within 6 h. The in-vitro simulation of in-vivo plasma and tissue pharmacokinetics of cefpirome and fosfomycin has shown that both antimicrobial agents kill S. aureus and P. aeruginosa strains effectively after single dose administration. This effect was most pronounced by the combined use of these antimicrobial agents. Therefore, this data corroborates antimicrobial strategies of simultaneous administration of cefpirome and fosfomycin in patients with severe soft tissue infection.

Impact of path parameters on maze solution time.

In order to compare spatial attention and visual processing capabilities of humans and rhesus macaques, we developed a visual maze task both could perform. Maze stimuli were constructed of orthogonal line segments displayed on a monitor. Each was octagonal in outline and contained a central square (the 'start box'). A single ('main') path, containing a random number of turns, extended outward from the start box, and either reached an exit in the maze's perimeter, or a blind ending within the maze. Subjects maintained ocular fixation within the start box, and indicated their judgment whether the path reached an exit or not by depressing one of two keys (humans) or foot pedals (monkeys). Successful maze solution by human subjects required a minimum viewing time. Replacing the maze with a masking stimulus after a variable interval revealed that the percent correct performance increased systematically with greater viewing time, reaching a plateau of approximately 85% correct if mazes were visible for 500 ms or more. A multiple linear regression analysis determined that the response time of both species depended upon several parameters of the main path, including the number of turns, total length, and exist status. Human and nonhuman primates required comparable time to process each turn in the path, whereas monkeys were faster than humans in processing each unit of path length. The data suggest that a covert analysis of the maze proceeds from the center outward along the main path in the absence of saccadic eye movements, and that both monkeys and humans undertake such an analysis during the solution of visual mazes.