Ration particle size has different effects on digestive but not production parameters in higher-yielding (Holstein) compared to lower-yielding (Girolando) cows.

The aim of the study was to evaluate the effect of total mixed ration particle size (length) and breed of cow on intake dynamics, animal performance and CH4 emissions, comparing high yielding Holstein and low yielding Girolando cows. The experimental design was 2 × 2 Latin Square arranged as a crossover factorial scheme with two diets (short particle size, SPS and long particle size, LPS) and the two breed compositions. The design comprised two periods of 26 d each, where all data collection was performed at cow level. No influence of the particle size occurred for the passage rate, neutral detergent fiber digestibility, performance and milk composition, methane emissions or ruminal fermentation parameters. Girolando cows had greater dry matter intake (DMI) when fed SPS, while Holsteins had the same (P < 0.05). Girolando cows had lower dry matter digestibility when fed LPS compared to SPS, while Holsteins had the opposite effect (P < 0.05). Also, the digestibility of crude protein and non-fibrous carbohydrates decreased in Girolando cows fed LPS, but not in Holsteins (P < 0.05). Girolando cows reduced DMI by 10.6% when fed LPS diet (P < 0.05). Girolando had an increased eating rate (+24 g of DM/min; P < 0.05) compared to Holstein cows, but Holstein cows had a lower CH4 intensity (by 29.7%: P < 0.05). Girolando cows increased the dry matter intake when fed a diet with short particle size, while the same did not happen in Holsteins. Dry matter digestibility increased in Holsteins when fed long particle size, while the opposite was observed in Girolando cows. Nutrient digestibility was reduced in Girolando cows when fed short particle size. Particle size did not influence eating time, eating rate, feed trough visits, visits with intake, milk yield and composition regardless of the breed. Reducing particle size increased CH4 intensity in both breeds.

Beyond Backpropagation: Bilevel Optimization Through Implicit Differentiation and Equilibrium Propagation.

This review examines gradient-based techniques to solve bilevel optimization problems. Bilevel optimization extends the loss minimization framework underlying statistical learning to systems that are implicitly defined through a quantity they minimize. This characterization can be applied to neural networks, optimizers, algorithmic solvers, and even physical systems and allows for greater modeling flexibility compared to the usual explicit definition of such systems. We focus on solving learning problems of this kind through gradient descent, leveraging the toolbox of implicit differentiation and, for the first time applied to this setting, the equilibrium propagation theorem. We present the mathematical foundations behind such methods, introduce the gradient estimation algorithms in detail, and compare the competitive advantages of the different approaches.

Adoption of Precision Technologies by Brazilian Dairy Farms: The Farmer's Perception.

The use of precision farming technologies, such as milking robots, automated calf feeders, wearable sensors, and others, has significantly increased in dairy operations over the last few years. The growing interest in farming technologies to reduce labor, maximize productivity, and increase profitability is becoming noticeable in several countries, including Brazil. Information regarding technology adoption, perception, and effectiveness in dairy farms could shed light on challenges that need to be addressed by scientific research and extension programs. The objective of this study was to characterize Brazilian dairy farms based on technology usage. Factors such as willingness to invest in precision technologies, adoption of sensor systems, farmer profile, farm characteristics, and production indexes were investigated in 378 dairy farms located in Brazil. A survey with 22 questions was developed and distributed via Google Forms from July 2018 to July 2020. The farms were then classified into seven clusters: (1) top yield farms; (2) medium-high yield, medium-tech; (3) medium yield and top high-tech; (4) medium yield and medium-tech; (5) young medium-low yield and low-tech; (6) elderly medium-low yield and low-tech; and (7) low-tech grazing. The most frequent technologies adopted by producers were milk meters systems (31.7%), milking parlor smart gate (14.5%), sensor systems to detect mastitis (8.4%), cow activity meter (7.1%), and body temperature (7.9%). Based on a scale containing numerical values (1-5), producers indicated "available technical support" (mean; σ2) (4.55; 0.80) as the most important decision criterion involved in adopting technology, followed by "return on investment-ROI" (4.48; 0.80), "user-friendliness" (4.39; 0.88), "upfront investment cost" (4.36; 0.81), and "compatibility with farm management software" (4.2; 1.02). The most important factors precluding investment in precision dairy technologies were the need for investment in other sectors of the farm (36%), the uncertainty of ROI (24%), and lack of integration with other farm systems and software (11%). Farmers indicated that the most useful technologies were automatic milk meters systems (mean; σ2) (4.05; 1.66), sensor systems for mastitis detection (4.00; 1.57), automatic feeding systems (3.50; 2.05), cow activity meter (3.45; 1.95), and in-line milk analyzers (3.45; 1.95). Overall, the concerns related to data integration, ROI, and user-friendliness of technologies are similar to those of dairy farms located in other countries. Increasing available technical support for sensing technology can have a positive impact on technology adoption.

A deep learning framework for neuroscience.

Systems neuroscience seeks explanations for how the brain implements a wide variety of perceptual, cognitive and motor tasks. Conversely, artificial intelligence attempts to design computational systems based on the tasks they will have to solve. In artificial neural networks, the three components specified by design are the objective functions, the learning rules and the architectures. With the growing success of deep learning, which utilizes brain-inspired architectures, these three designed components have increasingly become central to how we model, engineer and optimize complex artificial learning systems. Here we argue that a greater focus on these components would also benefit systems neuroscience. We give examples of how this optimization-based framework can drive theoretical and experimental progress in neuroscience. We contend that this principled perspective on systems neuroscience will help to generate more rapid progress.

Relationship between feed efficiency indexes and performance, body measurements, digestibility, energy partitioning, and nitrogen partitioning in pre-weaning dairy heifers.

The objectives of this study were: 1) to classify animals into groups of high and low feed efficiency using two feed efficiency indexes (Residual feed intake (RFI) and residual feed intake and body weight gain (RIG)), and 2) to evaluate if pre-weaning heifer calves divergent for feed efficiency indexes exhibit differences in performance, body measurements, digestibility, energy partitioning, and nitrogen partitioning. A total of 32 Gyr heifer calves were enrolled in a 63-d trial and classified into two feed efficiency (FE) groups based on RFI and RIG (mean ± 0.5 SD). The groups were classified as high efficiency (HE) RFI (HE RFI, n = 9; HE RIG, n = 10), and low efficiency (LE) RFI (LE RFI, n = 10; LE RIG, n = 11). The remaining animals were classified as intermediate (n = 13 (RFI) and n = 11 (RIG)). HE and LE calves had RFI values of-0.052 and 0.049 kg/d (P < 0.05), respectively. The HE RFI group consumed 8.9% less solid diet than the LE RFI group. HE RFI animals exhibited an increased digestibility of crude protein and ether extract and tended to have greater total dry and organic matter digestibility. LE RFI animals had greater gross energy and nitrogen intake, though greater fecal losses resulted in a tendency to reduce energy and nitrogen use efficiency. HE and LE calves had RIG values of 0.080 and -0.077kg/d (P ≤ 0.01), respectively. HE RIG animals exhibited greater average daily gain (9.4%), body weight (BW), and heart girth, though HE RIG group exhibited narrower hip width. HE RIG animals tended to have greater ether extract digestibility but greater methane losses (% of gross energy). HE RFI in pre-weaning heifers seems to be related to differences in digestibility. Divergent animals for RIG during the assessed phase appear to differ in body measurements, which may be related to differences in the composition of the gain.

Replacement of Soybean Meal With Soybean Cake Reduces Methane Emissions in Dairy Cows and an Assessment of a Face-Mask Technique for Methane Measurement.

The objective of this study was to (a) evaluate the effect of replacing soybean meal (SBM) with soybean cake (SBC) on feeding behavior, rumen fermentation, milk production, nutrient digestibility and CH4 emissions and (b) investigate whether a face-mask technique could be used to predict daily methane (CH4) emissions in dairy cattle. The experiment was conducted as a completely randomized design, with 32 crossbred Holstein × Gyr cows (days in milk (DIM): 112 ± 25.1) randomly assigned to the following treatments (n = 8/group) for 75 days: (1) 0% SBC, (2) 6% SBC, (3) 14% SBC, and (4) 23% SBC, in place of SBM on a dry matter (DM) basis. Across the final 4 weeks of the study, CH4 production was estimated using the proposed face-mask technique subsequent to a respiration chamber measurement for an evaluation of treatment efficacy and face-mask accuracy. There was no effect of SBM replacement by SBC on intake, feeding or drinking behavior (P > 0.21). Total VFA concentration, the individual proportions of VFA and blood metabolites were not altered (P > 0.17) by SBC, however there was a tendency for decreased (P = 0.08) lactate and plasma urea nitrogen (P = 0.07) concentration associated with SBC addition. Fat-corrected milk yield (FCM4%) and composition was not affected (P > 0.27) by SBC; however, there was a tendency for decreased total milk solids (P = 0.07) and milk fat (P = 0.08) associated with 23% SBC treatment. There was no treatment × technique interaction (P > 0.05) effect on gas measurements. A maximum reduction (P = 0.01) in CH4 yield (g/kg DM) and intensity (g/kg milk) of 11 and 20%, respectively, was observed for the 14% SBC inclusion. Compared to the week of mask measurements, chambers decreased (P = 0.01) intake (kg/d, %BW) and increased (P = 0.05) FCM4%. The face-mask method over estimated O2 consumption by 5%. The face-mask method accurately predicted daily CH4 emissions when compared to the chamber at the same time-point. However, there was a linear bias of CH4 outputs so further evaluation of the calculation of total CH4 from a spot measurement is required.

Computational roles of plastic probabilistic synapses.

The probabilistic nature of synaptic transmission has remained enigmatic. However, recent developments have started to shed light on why the brain may rely on probabilistic synapses. Here, we start out by reviewing experimental evidence on the specificity and plasticity of synaptic response statistics. Next, we overview different computational perspectives on the function of plastic probabilistic synapses for constrained, statistical and deep learning. We highlight that all of these views require some form of optimisation of probabilistic synapses, which has recently gained support from theoretical analysis of long-term synaptic plasticity experiments. Finally, we contrast these different computational views and propose avenues for future research. Overall, we argue that the time is ripe for a better understanding of the computational functions of probabilistic synapses.

Sensory representation of an auditory cued tactile stimulus in the posterior parietal cortex of the mouse.

Sensory association cortices receive diverse inputs with their role in representing and integrating multi-sensory content remaining unclear. Here we examined the neuronal correlates of an auditory-tactile stimulus sequence in the posterior parietal cortex (PPC) using 2-photon calcium imaging in awake mice. We find that neuronal subpopulations in layer 2/3 of PPC reliably represent texture-touch events, in addition to auditory cues that presage the incoming tactile stimulus. Notably, altering the flow of sensory events through omission of the cued texture touch elicited large responses in a subset of neurons hardly responsive to or even inhibited by the tactile stimuli. Hence, PPC neurons were able to discriminate not only tactile stimulus features (i.e., texture graininess) but also between the presence and omission of the texture stimulus. Whereas some of the neurons responsive to texture omission were driven by looming-like auditory sounds others became recruited only with tactile sensory experience. These findings indicate that layer 2/3 neuronal populations in PPC potentially encode correlates of expectancy in addition to auditory and tactile stimuli.

Tithonia diversifolia as a Supplementary Feed for Dairy Cows.

The objective of this study was to examine the effects of Tithonia diversifolia as a supplementary forage on dairy cow performance and methane production. Nine lactating Holstein × Zebu dairy cows (519 ± 53.3 kg of body weight and 66 ± 13.3 d in milk) were paired by milk yield (21.3 ± 2.34 kg/d) and body weight and randomly assigned to three dietary treatments in a Latin square design with 21-d experimental periods (14 d for diet adaptation and 7 d for measurements and sample collection). The dietary treatments included the control diet consisting of fresh sugar cane plus concentrate (44:56, % of diet DM), and two treatment diets containing different levels of fresh T. diversifolia (6.5 and 15.4%, DM basis) which partially replaced both sugarcane and concentrates. Methane production was measured using the sulphur hexafluoride (SF6) technique from d 16 to d 21 of each experimental period. Analysis of the gas samples was performed by gas chromatography. The inclusion of T. diversifolia at 15.4% DM had no effects on DM intake, milk production, nitrogen balance or methane production. There was no effect on the concentrations of total saturated fatty acids (SFA), monounsaturated fatty acids (MUFA), and polyunsaturated fatty acids (PUFA) in milk fat (P ≥ 0.28), though individual milk fatty acids were affected. Serum concentrations of glucose, urea nitrogen (BUN), triglycerides, ß-hydroxybutyrate (BHBA), and cholesterol were unaffected by the dietary treatments (P ≥ 0.13). There was a time (2 and 6 h post-feeding) and dietary treatment effect (P < 0.01) on the acetate to propionate ratio in the rumen. A denaturing gradient gel electrophoresis analysis of the archaeal community showed distinct clustering of the archaea populations for control and treatment diets. Taken together, our results indicate the potential of T. diversifolia as a supplementary forage for dairy cattle in the tropics.

Productivity and nutritional quality of Flechinha grass ( Echinolaena inflexa ), native grass of Brazilian Cerrado / Produção e qualidade nutricional do capim-flechinha (Echinolaena inflexa), gramínea nativa do Cerrado brasileiro

ABSTRACT: Due to scarce nutritional data, this study assessed the productivity and nutritional value of Echinolaena inflexa (EI) grass, native to the Cerrado biome. It was compared to B. brizantha (BB), one of the most cultivated grasses in Brazil, during a whole year (rainy; RS and dry season; DS). Sampling was held in accordance with pasture management (entry / exit height; 50 / 5cm and 80 / 25cm for EI and BB, respectively). Dry matter production (DMP), crude protein (CP), neutral and acid detergent fiber (NDF; ADF), hemicellulose (HCEL), PB insoluble neutral and acid detergent (PIDN; PIDA), total and non-fibrous carbohydrates (TC; NFC), ether extract (EE), and mineral matter (MM), and in vitro fermentation kinetics and DM degradability (DMD) were evaluated. A completely randomized design (season as a fixed term) and average treatment compared by Tukey post test were applied. EI produced 38.5% of the DMP of BB. A higher CP (75.3; 73.5 in the RS and DS), PIDA (12.5; 8.7), PIDN (47.1; 40.1), NDF (714.4; 749.5) and ADF (396.0; 419.0) were obtained by EI in relation to BB (CP (60.3; 33.5), PIDA (6.0; 3.5), PIDN (21.4; 10.8), NDF (673.0; 675.1) and ADF (335.5; 351.4) during the RS and DS, respectively (g kg-1 DM). In vitro data were directly associated with chemical composition, resulting in lower DMD of EI compared to BB. EI showed productive similarity (DMP) during RS and DS (939.3; 809.8kg DM respectively). Although EI showed greater nutritional stability (CP) between seasons, 17% of CP was linked to ADF and therefore, not available for rumen microorganisms.

RESUMO: Existem escassas informações a respeito da qualidade nutricional e potencial produtivo de Echinolaena Inflexa, planta nativa do Cerrado Brasileiro e que possui potencial valor forrageiro. Dessa forma, este trabalho objetivou avaliar o seu valor nutritivo e produtividade, comparando-a com a B. brizantha (BB), uma das mais cultivadas gramíneas no Brasil, durante um no agrícola (Estação das Chuvas; CH e seca; SE). A coleta do material experimental foi realizada de acordo com as recomendações de manejo (altura de entrada / altura de saída 50/5cm e 80/25cm para EI e BB, respectivamente). Foram avaliadas a produção de matéria seca (PMS), proteína bruta (PB), fibra detergente neutro e ácido (FDN e FDA), hemicelulose (HCEL), PB insolúvel em detergente neutro e ácido (PIDN e PIDA), carboidratos totais e não fibrosos (CT e CNF), extrato etéreo (EE) e matéria mineral (MM), e a cinética de fermentação ruminal e degradabilidade in vitro da MS (DMS). Utilizou-se o delineamento inteiramente casualizado e as médias de tratamento comparadas pelo post test de Tukey. A EI apresentou 38,5% da PMS da BB. Foram obtidos maiores teores de PB (75,3 e 73,5), PIDA (12,5 e 8,7), PIDN (47,1 e 40,1), FDN (714,4 e 749,5) e FDA (396,0 e 419,0) para a EI em relação a BB (PB; 60,3 e 33,5), PIDA (6,0 e 3.5), PIDN (21,4 e 10,8), FDN (673,0 e 675,1) e FDA (335,5 e 351,4) na estação chuvosa e seca, respectivamente (g kg-1 MS). Os dados in vitro da EI foram diretamente associados à composição química, resultando em menor DMS comparativamente BB. A EI apresentou estabilidade produtiva (PMS) durante a estação chuvosa e seca (939,3 e 809,8kg MS), respectivamente. Apesar da similaridade nutricional (PB) nas estações, cerca de 17% da PB está ligada a FDA e, portanto, não disponível para os microrganismos ruminais.

Surveillance of surgical site infection after cesarean section and time of notification.

BACKGROUND: Cesarean section is a surgical procedure the main complication of which is surgical site infection (SSI), which is related to maternal morbidity and mortality. OBJECTIVE: To evaluate active monitoring by telephone to identify infection and time of SSI report in postpartum women and associated risk factors. METHODS: We conducted a prospective observational study from 2013-2014, at a referral service for high-risk pregnancies. Surveillance was conducted via telephone at least 30 days after cesarean delivery. Incidence ratio and time of infection occurrence (days) was analyzed. Survival analysis was conducted to assess the temporal distribution of the development of infection. RESULTS: Of a total of 353 patients, 14 (4%) cases of SSI were reported, and 10 (7.4%) of the reported cases occurred within 15 days after cesarean and average time of infection was12.21 days. American Society of Anesthesiologists score was the only risk factor associated with SSI after cesarean section. CONCLUSIONS: The prevalence of SSI after cesarean section via telephone is similar to several services with different methods of surveillance, considering it could be used by services with limited resources. Superficial incisional SSI was the most common type of infection, time of infection report was mainly before the 15th day postprocedure, and American Society of Anesthesiologists score of 2 or less was protective against SSI. Telephone calls can be a viable method to identify women with infection briefly after discharge, particularly at-risk patients.

Estudo transversal do perfil das infecções congênitas atendidas entre 2012 e 2014 em um centro de referência em Belo Horizonte, Minas Gerais / Cross-sectional study of clinical and laboratorial aspects of congenital infections attended at a reference Center in Belo Horizonte, MG, 2012-2014

Introdução: as infecções congênitas são resultantes da transmissão vertical de microrganismos de gestantes infectadas para seus conceptos. Apesar dessas infecções, em geral, cursarem com pouca ou nenhuma manifestação clínica nas gestantes, a infecção fetal pode trazer morbimortalidade perinatal e na infância. Objetivo: identificar a prevalência das infecções congênitas encaminhadas ao Centro de Referência e Treinamento em Doenças Infecciosas e Parasitárias Orestes Diniz (CTR/DIP Orestes Diniz) e avaliar os métodos laboratoriais usados para o diagnóstico. Métodos: estudo transversal realizado em ambulatório de referência em doenças infecciosas, a partir de coleta de dados de prontuários de crianças com diagnóstico suspeito de infecção congênita. A confirmação diagnóstica baseou-se em testes sorológicos ou de biologia molecular, além de descrição de sintomatologia da criança. Resultados: um total de 513 crianças foram identificadas, sendo que 41,3% tiveram o diagnóstico confirmado, a maioria foi de toxoplasmose (45,35%) e sífilis (15,98%). Entre as crianças com diagnóstico confirmado, 28,85% apresentaram manifestações clínicas, enquanto que no grupo com diagnóstico indeterminado ou suspeito o percentual foi de 16,38%. As principais manifestações identificadas foram acometimento do sistema nervoso central (n=39) e alterações visuais (n=30). Conclusão: a confirmação de infecção foi definida em aproximadamente metade dos pacientes avaliados e a maioria das crianças foi assintomática ao nascimento. O pré-natal de qualidade e a propedêutica e tratamento precoce das crianças identificadas podem reduzir o impacto dessas infecções no nosso meio.(AU)

Introduction: congenital infections are results of microorganisms transmitted to the fetus by the infected pregnant. Most newborn infants infected during pregnancy or labor have no sings of congenital disease. However, these infections may cause perinatal and infancy morbidity and mortality. Objective: to determine the prevalence of congenital infections in newborns and infants attended at the CTR/DIP Orestes Diniz (Centro de Treinamento e Referência em Doenças Infecciosas e Parasitárias Orestes Diniz) and to analyse the laboratorial methods used for diagnosis of congenital disease of assisted children. Methods: cross-sectional study conducted in an Infectious Diseases Reference Center where it was evaluated the charts of infants with suspected congenital infection. Diagnosis confirmation was based on serological tests, molecular biology and signs and symptoms described in the charts. Results: A total of 41,3% of the 513 children identified have had a defined diagnosis. Most of them had toxoplasmosis (45,35%) and syphilis (15,98%). Clinical manifestations was observed in 28,85% of children with defined diagnosis of congenital infection, and in 16,38% of children with uncertain diagnosis. Central Nervous System (n=39) and ocular (n=30) manifestations were the most frequent findings. Conclusions: Defined diagnosis was possible in about half of cases and most of them were asymptomatic at birth. An appropriate prenatal care and early diagnosis and treatment of congenital infections may reduce the impact of disease in the population.(AU)

Management of syphilis in pregnant womenand their newborns: is it still a problem? / Manejo de sífilis em gestantes e seus recém-nascidos: ainda um problema?

A sífilis congênita (SC) é um agravo prevenível, mas o Brasil ainda apresenta alta prevalência da doença, com consequente morbimortalidade perinatal. Objetivo: Avaliar a abordagem de sífilis em gestantes e seus recém-nascidos encaminhados para centro de referência. Métodos: Estudo transversal, de março de 2012 a abril de 2013. A coleta de dados foi realizada em prontuários de pacientes referenciados com SC, considerando critérios estabelecidos pelo Ministério da Saúde (MS). Os dados foram analisados pelo Statistical Package for the Social Sciences (SPSS) e o estudo foi aprovado pelo Comitê de Ética. Resultados: Um total de 31 recém-nascidos foi encaminhado devido à triagem materna com Venereal Disease Research Laboratory(VDRL) materno positivo durante a gestação, com 4 mulheres adequadamente tratadas. Treze recém-nascidos apresentaram alteração no hemograma e1 apresentou alteração óssea, 28 deles com tratamento adequado. Discussão: Quando se considera adequação de tratamento de acordo com as diretrizes nacionais, poucos casos de sífilis na gestação são considerados adequadamente tratados. Isso impacta na assistência ao recém-nascido, que, muitas vezes,é submetido a propedêutica invasiva e tratamento extenso, embora na maioria das vezes seja assintomático. Conclusão: O seguimento das recomendações para o tratamento da sífilis na gestante tem sido, frequentemente, considerado inadequado, o que dificulta a eliminação da SC.

Congenital syphilis (CS) is a preventable disease, but its prevalence is still high in Brazil, with consequent perinatal morbidity and mortality.Objective: To evaluate the approach of syphilis in pregnant women and their newborns referred to the referral center of Orestes Diniz, in Belo Horizonte.Methods: A cross-sectional study was carried out from March 2012 to April 2013. Data collection was performed on the medical records of patients referred with CS, considering the criteria established by the Ministry of Health. Data were analyzed using SPSS and the study was approved by the Ethics Committee. Results: A total of 31 newborns were referred due to a positive result in maternal testing with Venereal Disease Research Laboratory during pregnancy. However, only four women have been adequately treated in accordance with the Ministry of Health. Thirteen newborns presented alterations inblood cells count, one had bone rarefactions, and 28 presented proper information of treatment. Discussion: When considering the adequacy of treatmenta ccording to the national guidelines, few cases of syphilis during pregnancy can be considered adequately treated. This affects the assistance to the newborn,who is often subjected to invasive investigation and extensive treatment, although most are asymptomatic. Conclusion: The follow-up of recommendations for the treatment of syphilis in pregnant women has often been considered inadequate, making CS difficult to eliminate

Energy Efficient Sparse Connectivity from Imbalanced Synaptic Plasticity Rules.

It is believed that energy efficiency is an important constraint in brain evolution. As synaptic transmission dominates energy consumption, energy can be saved by ensuring that only a few synapses are active. It is therefore likely that the formation of sparse codes and sparse connectivity are fundamental objectives of synaptic plasticity. In this work we study how sparse connectivity can result from a synaptic learning rule of excitatory synapses. Information is maximised when potentiation and depression are balanced according to the mean presynaptic activity level and the resulting fraction of zero-weight synapses is around 50%. However, an imbalance towards depression increases the fraction of zero-weight synapses without significantly affecting performance. We show that imbalanced plasticity corresponds to imposing a regularising constraint on the L1-norm of the synaptic weight vector, a procedure that is well-known to induce sparseness. Imbalanced plasticity is biophysically plausible and leads to more efficient synaptic configurations than a previously suggested approach that prunes synapses after learning. Our framework gives a novel interpretation to the high fraction of silent synapses found in brain regions like the cerebellum.

Binary Willshaw learning yields high synaptic capacity for long-term familiarity memory.

In this study, we investigate from a computational perspective the efficiency of the Willshaw synaptic update rule in the context of familiarity discrimination, a binary-answer, memory-related task that has been linked through psychophysical experiments with modified neural activity patterns in the prefrontal and perirhinal cortex regions. Our motivation for recovering this well-known learning prescription is two-fold: first, the switch-like nature of the induced synaptic bonds, as there is evidence that biological synaptic transitions might occur in a discrete stepwise fashion. Second, the possibility that in the mammalian brain, unused, silent synapses might be pruned in the long-term. Besides the usual pattern and network capacities, we calculate the synaptic capacity of the model, a recently proposed measure where only the functional subset of synapses is taken into account. We find that in terms of network capacity, Willshaw learning is strongly affected by the pattern coding rates, which have to be kept fixed and very low at any time to achieve a non-zero capacity in the large network limit. The information carried per functional synapse, however, diverges and is comparable to that of the pattern association case, even for more realistic moderately low activity levels that are a function of network size.

Regarding the temporal requirements of a hierarchical Willshaw network.

In a recent communication, Sacramento and Wichert (2011) proposed a hierarchical retrieval prescription for Willshaw-type associative networks. Through simulation it was shown that one could make use of low resolution descriptor patterns to decrease the total time requirements of recalling a learnt association. However, such a method introduced a dependence on a set of new parameters which define the structure of the hierarchy. In this work we compute the expected retrieval time for the random neural activity regime which maximises the capacity of the Willshaw model and we study the task of finding the optimal hierarchy parametrisation with respect to the derived temporal expectation. Still in regard to this performance measure, we investigate some asymptotic properties of the algorithm.

Tree-like hierarchical associative memory structures.

In this letter we explore an alternative structural representation for Steinbuch-type binary associative memories. These networks offer very generous storage capacities (both asymptotic and finite) at the expense of sparse coding. However, the original retrieval prescription performs a complete search on a fully-connected network, whereas only a small fraction of units will eventually contain desired results due to the sparse coding requirement. Instead of modelling the network as a single layer of neurons we suggest a hierarchical organization where the information content of each memory is a successive approximation of one another. With such a structure it is possible to enhance retrieval performance using a progressively deepening procedure. To backup our intuition we provide collected experimental evidence alongside comments on eventual biological plausibility.