Hebbian dreaming for small datasets

The dreaming Hopfield model constitutes a generalization of the Hebbian paradigm for neural networks, that is able to perform on-line learning when "awake" and also to account for off-line "sleeping" mechanisms. The latter have been shown to enhance storing in such a way that, in the long sleep-time limit, this model can reach the maximal storage capacity achievable by networks equipped with symmetric pairwise interactions. In this paper, we inspect the minimal amount of information that must be supplied to such a network to guarantee a successful generalization, and we test it both on random synthetic and on standard structured datasets (i.e., MNIST, Fashion-MNIST and Olivetti). By comparing these minimal thresholds of information with those required by the standard (i.e., always "awake") Hopfield model, we prove that the present network can save up to ∼90% of the dataset size, yet preserving the same performance of the standard counterpart. This suggests that sleep may play a pivotal role in explaining the gap between the large volumes of data required to train artificial neural networks and the relatively small volumes needed by their biological counterparts. Further, we prove that the model Cost function (typically used in statistical mechanics) admits a representation in terms of a standard Loss function (typically used in machine learning) and this allows us to analyze its emergent computational skills both theoretically and computationally: a quantitative picture of its capabilities as a function of its control parameters is achieved and consistency between the two approaches is highlighted. The resulting network is an associative memory for pattern recognition tasks that learns from examples on-line, generalizes correctly (in suitable regions of its control parameters) and optimizes its storage capacity by off-line sleeping: such a reduction of the training cost can be inspiring toward sustainable AI and in situations where data are relatively sparse.

Quantifying heterogeneity to drug response in cancer-stroma kinetics

A crucial challenge in medicine is choosing which drug (or combination) will be the most advantageous for a particular patient. Usually, drug response rates differ substantially, and the reasons for this response unpredictability remain ambiguous. Consequently, it is central to classify features that contribute to the observed drug response variability. Pancreatic cancer is one of the deadliest cancers with limited therapeutic achievements due to the massive presence of stroma that generates an environment that enables tumor growth, metastasis, and drug resistance. To understand the cancer-stroma cross talk within the tumor microenvironment and to develop personalized adjuvant therapies, there is a necessity for effective approaches that offer measurable data to monitor the effect of drugs at the single-cell level. Here, we develop a computational approach, based on cell imaging, that quantifies the cellular cross talk between pancreatic tumor cells (L3.6pl or AsPC1) and pancreatic stellate cells (PSCs), coordinating their kinetics in presence of the chemotherapeutic agent gemcitabine. We report significant heterogeneity in the organization of cellular interactions in response to the drug. For L3.6pl cells, gemcitabine sensibly decreases stroma-stroma interactions but increases stroma-cancer interactions, overall enhancing motility and crowding. In the AsPC1 case, gemcitabine promotes the interactions among tumor cells, but it does not affect stroma-cancer interplay, possibly suggesting a milder effect of the drug on cell dynamics.

Probing Single-Cell Fermentation Fluxes and Exchange Networks via pH-Sensing Hybrid Nanofibers

The homeostatic control of their environment is an essential task of living cells. It has been hypothesized that, when microenvironmental pH inhomogeneities are induced by high cellular metabolic activity, diffusing protons act as signaling molecules, driving the establishment of exchange networks sustained by the cell-to-cell shuttling of overflow products such as lactate. Despite their fundamental role, the extent and dynamics of such networks is largely unknown due to the lack of methods in single-cell flux analysis. In this study, we provide direct experimental characterization of such exchange networks. We devise a method to quantify single-cell fermentation fluxes over time by integrating high-resolution pH microenvironment sensing via ratiometric nanofibers with constraint-based inverse modeling. We apply our method to cell cultures with mixed populations of cancer cells and fibroblasts. We find that the proton trafficking underlying bulk acidification is strongly heterogeneous, with maximal single-cell fluxes exceeding typical values by up to 3 orders of magnitude. In addition, a crossover in time from a networked phase sustained by densely connected "hubs" (corresponding to cells with high activity) to a sparse phase dominated by isolated dipolar motifs (i.e., by pairwise cell-to-cell exchanges) is uncovered, which parallels the time course of bulk acidification. Our method addresses issues ranging from the homeostatic function of proton exchange to the metabolic coupling of cells with different energetic demands, allowing for real-time noninvasive single-cell metabolic flux analysis.

Outperforming RBM Feature-Extraction Capabilities by "Dreaming" Mechanism

Inspired by a formal equivalence between the Hopfield model and restricted Boltzmann machines (RBMs), we design a Boltzmann machine, referred to as the dreaming Boltzmann machine (DBM), which achieves better performances than the standard one. The novelty in our model lies in a precise prescription for intralayer connections among hidden neurons whose strengths depend on features correlations. We analyze learning and retrieving capabilities in DBMs, both theoretically and numerically, and compare them to the RBM reference. We find that, in a supervised scenario, the former significantly outperforms the latter. Furthermore, in the unsupervised case, the DBM achieves better performances both in features extraction and representation learning, especially when the network is properly pretrained. Finally, we compare both models in simple classification tasks and find that the DBM again outperforms the RBM reference.

Fully Automated Computational Approach for Precisely Measuring Organelle Acidification with Optical pH Sensors

pH balance and regulation within organelles are fundamental to cell homeostasis and proliferation. The ability to track pH in cells becomes significantly important to understand these processes in detail. Fluorescent sensors based on micro- and nanoparticles have been applied to measure intracellular pH; however, an accurate methodology to precisely monitor acidification kinetics of organelles in living cells has not been established, limiting the scope of this class of sensors. Here, silica-based fluorescent microparticles were utilized to probe the pH of intracellular organelles in MDA-MB-231 and MCF-7 breast cancer cells. In addition to the robust, ratiometric, trackable, and bioinert pH sensors, we developed a novel dimensionality reduction algorithm to automatically track and screen massive internalization events of pH sensors. We found that the mean acidification time is comparable among the two cell lines (ΔT_MCF-7 = 16.3 min; ΔT_MDA-MB-231 = 19.5 min); however, MCF-7 cells showed a much broader heterogeneity in comparison to MDA-MB-231 cells. The use of pH sensors and ratiometric imaging of living cells in combination with a novel computational approach allow analysis of thousands of events in a computationally inexpensive and faster way than the standard routes. The reported methodology can potentially be used to monitor pH as well as several other parameters associated with endocytosis.

Measurement of the Cosmic Ray Helium Energy Spectrum from 70 GeV to 80 TeV with the DAMPE Space Mission

The measurement of the energy spectrum of cosmic ray helium nuclei from 70 GeV to 80 TeV using 4.5 years of data recorded by the Dark Matter Particle Explorer (DAMPE) is reported in this work. A hardening of the spectrum is observed at an energy of about 1.3 TeV, similar to previous observations. In addition, a spectral softening at about 34 TeV is revealed for the first time with large statistics and well controlled systematic uncertainties, with an overall significance of 4.3σ. The DAMPE spectral measurements of both cosmic protons and helium nuclei suggest a particle charge dependent softening energy, although with current uncertainties a dependence on the number of nucleons cannot be ruled out.

Abstract 2967: Microgel-based in vitro tumoroid platform for real time assessment of drug sensitivity and resistance

While 3D cell cultures have been used as suitable representatives of in vivo conditions compared to 2D systems, scalability and flexibility in designing such platforms has been a major challenge1. Micron-sized 3D culture platforms offer the possibility of high throughput analysis, however major challenges exist with respect to their overall design. Technical obstacles include difficulties in supporting multiple cell types (e.g. co-cultures of cancer associated fibroblasts (CAFs) and tumor cells) as well as enabling monitoring of important parameters, such as proliferation and metabolism, in real time. Our 3D co-culture systems are based on pancreatic ductal adenocarcinoma (PDAC) spheroid models in microgels, generated by coaxial droplet microfluidics. This allows for co-culturing of cancer cells with the supporting cell types that constitute the major portion of the actual tumor. Initial optimization experiments utilized co-culture of commercial pancreatic cancer cell lines such, as L3.6pl cells, and stromal CAFs. In addition to optimizing the 3D microcultures using commercial PDAC cell lines, we are also modifying the system to take advantage of patient-matched CAFs and conditionally reprogrammed (CR) primary PDAC cells. Both drug sensitive and isogenic nab-paclitaxel resistant PDAC CR cells will be used, as recently published2. Our preliminary data showed that our platform supports rapid generation of numerous identical spheroids inside the microgel. Using microscopy under different treatment conditions (e.g the presence/absence of drugs) we can quantify spheroid growth. In addition, spatio-temporal mapping of the microenvironmental parameters such as dissolved O2, pH and K+ during growth is performed by time lapse fluorescence microscopy using fluorescent silica microparticles embedded in the matrix. This process will yield an in-depth understanding of the kinetics of growth and associated biochemical and metabolic activity for each cell type (e.g. CAF and PDAC) under different treatment conditions. Our T3D platform offers a rapid, quantifiable assessment of the behavior of patient-derived cells in providing additional details related to the changing microenvironment and may advance our understanding of treatment response and failures.

The research leading to these results received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement No 759959, ERC-StG “INTERCELLMED”).

Reference

(1) Nunes et al. Biotechnol. Bioeng. 2019, 116 (1), 206-226.

(2) Parasido et al. Mol. Cancer Res. 2019, 17 (9), 1815-1827.

Citation Format: Anil Chandra, Saumya Prasad, Francesco Alemanno, Adriano Barra, Enza Lonardo, Erika Parasido, Christopher Albanese, Loretta L. del Mercato. Microgel-based in vitro tumoroid platform for real time assessment of drug sensitivity and resistance [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2967.

Abstract A48: Quantifying stroma-tumor cell interactions in three-dimensional cell culture systems

Background: In cancer research, studying cell motility is fundamental to investigate cancer invasion and drug resistance. In solid tumors constituted by a huge stroma component, such as pancreatic ductal adenocarcinoma (PDAC), the ability to quantify cell movement and interactions is mandatory to better understand the complex crosstalk between cancer and stroma cells. In PDAC, the pancreatic stellate cells (PSCs) are the principal source of dense fibrotic stroma: these closely interact with the PDAC cells to create a facilitating tumor microenvironment that supports local and distant tumor progression through the secretion and/or the response to a number of cytokines that globally increase cancer invasiveness; moreover, the dense tumor microenvironment contributes to resistance to chemotherapy and radiation therapy. Recent studies have established that targeting the stromal compartment in PDAC may lead to promising outcomes. Unfortunately, significant improvements in the overall survival of patients have not been realized in more than four decades, in part because of the lack of relevant preclinical models.

Methods: To infer interactions existing between stromal and cancer cells in complex 3D environments, we developed a novel platform that combines time-lapse fluorescence microscopy, automated image detection, and extensive statistical analysis. To better mimic in situ PDAC, we created a hydrogel-based (Matrigel, collagen) model that allows cell tracking in 4 dimensions (x, y, z, t), not possible in 2D, using a mixed population of L3.6PL pancreatic cancer cells and PSCs at a ratio of 25% and 75%, respectively. To precisely identify the two populations and facilitate automatic cell detection, cells were transfected with GFP and mCherry vectors. Once in the 3D conformation, the cocultures were monitored through time-lapse confocal microscopy (CLSM) in controlled conditions, and the dataset containing the temporal evolution of the cells was processed by statistical tools.

Results: Key to our approach was the generation of new machine learning-driven automated inferential protocols that resulted in the high-resolution imaging of the strength of the interactions between cells, as well as the potential presence of local chemokine gradients. In our ongoing studies, both commercial PDAC cell lines and pancreatic stellate cells were used to establish the model system. We are currently adopting the model to take full advantage of patient-derived PDAC cells established using the conditionally reprogrammed cells technique, mixed with patient-matched stellate cells. Since cell migration is a hallmark of cancer, we believe that this platform could be used as a reliable and reproducible approach for studying single-cell migration and invasion also in patient-derived models. The research leading to these results received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement No. 759959, ERC-StG “INTERCELLMED”).

Citation Format: Marta Maria Cavo, Francesco Alemanno, Donatella Delle Cave, Eliana D'Amone, Adriano Barra, Enza Lonardo, Erika Parasido, Chris Albanese, Loretta Laureana del Mercato. Quantifying stroma-tumor cell interactions in three-dimensional cell culture systems [abstract]. In: Proceedings of the AACR Special Conference on the Evolving Landscape of Cancer Modeling; 2020 Mar 2-5; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2020;80(11 Suppl):Abstract nr A48.

Generalized Guerra's interpolation schemes for dense associative neural networks

In this work we develop analytical techniques to investigate a broad class of associative neural networks set in the high-storage regime. These techniques translate the original statistical-mechanical problem into an analytical-mechanical one which implies solving a set of partial differential equations, rather than tackling the canonical probabilistic route. We test the method on the classical Hopfield model - where the cost function includes only two-body interactions (i.e., quadratic terms) - and on the "relativistic" Hopfield model - where the (expansion of the) cost function includes p-body (i.e., of degree p) contributions. Under the replica symmetric assumption, we paint the phase diagrams of these models by obtaining the explicit expression of their free energy as a function of the model parameters (i.e., noise level and memory storage). Further, since for non-pairwise models ergodicity breaking is non necessarily a critical phenomenon, we develop a fluctuation analysis and find that criticality is preserved in the relativistic model.

Neural Networks with a Redundant Representation: Detecting the Undetectable

We consider a three-layer Sejnowski machine and show that features learnt via contrastive divergence have a dual representation as patterns in a dense associative memory of order P=4. The latter is known to be able to Hebbian store an amount of patterns scaling as N^{P-1}, where N denotes the number of constituting binary neurons interacting P wisely. We also prove that, by keeping the dense associative network far from the saturation regime (namely, allowing for a number of patterns scaling only linearly with N, while P>2) such a system is able to perform pattern recognition far below the standard signal-to-noise threshold. In particular, a network with P=4 is able to retrieve information whose intensity is O(1) even in the presence of a noise O(sqrt[N]) in the large N limit. This striking skill stems from a redundancy representation of patterns-which is afforded given the (relatively) low-load information storage-and it contributes to explain the impressive abilities in pattern recognition exhibited by new-generation neural networks. The whole theory is developed rigorously, at the replica symmetric level of approximation, and corroborated by signal-to-noise analysis and Monte Carlo simulations.

Peritonectomy and hyperthermic intraperitoneal chemotherapy: cost analysis and sustainability

BACKGROUND

Malignancies of the peritoneum remain a challenge in any hospital that accepts to manage them, due not only to difficulties associated with the complexity of the procedures involved but also the costs, which - in Italy and other countries that use a diagnosis-related group (DRG) system - are not adequately reimbursed.

MATERIAL AND METHODS

We analyzed data relative to 24 patients operated on between September 2010 and May 2013 with special regard to operating room expenditure, ICU stay, duration of hospitalization, and DRG reimbursement. The total costs per patient included clinical, operating room, procedure, pathology, imaging, ward care, allied healthcare, pharmaceutical, and ICU costs.

RESULTS

Postoperative hospital stay, drugs and materials, and operating room occupancy were the main factors affecting the expenditure for cytoreductive surgery and hyperthermic intraperitoneal chemotherapy. We had a median hospitalization of 14 days, median ICU stay of 2.4 days, and median operating room occupancy of 585 min. The median expenditure for each case was € 21,744; the median reimbursement by the national health system € 8,375.

CONCLUSIONS

In a DRG reimbursement system, the economic effort in the management of patients undergoing peritonectomy procedures may not be counterbalanced by adequate reimbursement. Joint efforts between medical and administration parties are mandatory to develop appropriate treatment protocols and keep down the costs.

Buprenorphine versus tramadol as perineural adjuvants for postoperative analgesia in patients undergoing arthroscopic rotator cuff repair under middle interscalene block: a retrospective study

BACKGROUND

The aim of this retrospective study was to compare buprenorphine and tramadol, in order to assess their different efficacy in prolonging postoperative analgesia and their associated side effects when used as perineural adjuvants with a local anesthetic.

METHODS

The clinical records of 161 consecutive ASA 1-2 adult patients scheduled for arthroscopic rotator cuff repair and fulfilling the inclusion/exclusion criteria were reviewed retrospectively. The anaesthesia was performed using the middle interscalene block (MIB). The 161 patients were divided into three groups (A, B, T) according to their utilization of buprenorphine (B), tramadol (T) or neither of the latter (A) as perineural adjuvants: group A (54 patients) - levobupivacaine 0.75%, 0.4 mL/kg; group B (56 patients) - levobupivacaine 0.75%, 0.4 mL/kg + 0.15 mg buprenorphine; group T (51 patients) - levobupivacaine 0.75%, 0.4 mL/kg + 100 mg tramadol.

RESULTS

The results showed that the group treated with buprenorphine benefited from a longer post-operative analgesia than that treated with local anesthetic alone (P<0.0001). Otherwise, a less evident not statistically significant (P=0.4825) difference turned out between the group treated with the anesthetic alone and the group treated with tramadol as adjuvant. No difference turned out to be between the local anesthetic alone treatment and the tramadol-local anesthetic one (P=0.4825; HR=0.863, 95% CI 0.574-1.299); on the contrary, a significant difference was demonstrated between the buprenophine-local anesthetic group and the local anesthetic alone one (P<0.0001; HR=0.330, 95% CI 0.216-0.530) CONCLUSION: Both buprenorphine and tramadol are effective as perineural adjuvants used in order to prolong the postoperative analgesia, buprenorphine proving more efficacious for this purpose than tramadol.

Tramadol and 0.5% levobupivacaine for single-shot interscalene block: effects on postoperative analgesia in patients undergoing shoulder arthroplasty

BACKGROUND

The aim of this study was to evaluate the efficacy of tramadol as an adjuvant to the local anaesthetic solution in patients undergoing shoulder arthroscopy for rotator cuff tear after middle interscalene block (MIB).

METHODS

We enrolled 120 patients (ASA I-II), scheduled for arthroscopic surgery for rotator cuff tear. The patients were sedated with midazolam 0.02 mg/kg and haloperidol 2 mg i.v. before performing MIB. All subjects underwent a MIB with 0.4 mL/kg of 0.5% levobupivacaine. After computerized randomization, all patients were allocated in 1 of 3 groups, each including 40 subjects. Group Placebo (Group P) received 0.4 mL/kg of 0.5% levobupivacaine plus isotonic sodium chloride for MIB and isotonic sodium chloride i.m. Group "Perineural Tramadol" (Group TPN) received 0.4 ml/Kg of 0.5% levobupivacaine plus 1.5 mg/kg of tramadol perineurally and isotonic sodium chloride i.m. Group "Intramuscular Tramadol" (Group TIM) received 0.4 ml/Kg of 0.5% levobupivacaine plus isotonic sodium chloride perineurally and 1.5 mg/kg of tramadol i.m.

RESULTS

The MIB onset times were not statistically different in the three groups. The duration of analgesia was significantly longer in Groups TPN and TIM, where tramadol was administered, either i.m. or perineurally, compared with the placebo group. A significant statistical difference was found in the duration of analgesia between the group TPN and TIM.

CONCLUSION

The addition of tramadol to the local anaesthetic solution administered for MIB provided a longer duration of analgesia compared with placebo and i.m tramadol administration in patients undergoing arthroscopic surgery for rotator cuff tear.

Alemanno's brachial plexus block ten years later: topographic study of the anesthetized areas

AIM

The study analyzes the extension of anesthesia induced by Alemanno's brachial plexus block technique on the various areas of competence of the different nerves.

METHODS

The study was conducted on 58 patients in ASA classes 1 and 2 scheduled to receive shoulder arthroscopy. At the end of the operation, about 2 hours after induction of anesthesia, extension of anesthesia was evaluated by the pin-prick test.

RESULTS

Anesthesia was achieved in 100% of cases for the circumflexus, musculocutaneous and radial nerves; the median nerve escaped in 7% of cases, the medial cutaneous nerve of forearm in 20%, the ulnar nerve in 27.5%, the medial cutaneous nerve of arm and intercostobrachial nerves in 46%; no major complications were associated with the technique.

CONCLUSION

Alemanno's technique is recommended for operations on the shoulder and humerus, whereas more peripheral techniques may be preferable for interventions on the forearm and hand.

[Sopraclavicular techniques for brachial plexus block]

BACKGROUND

Aim of the present study is to review the four different methods of brachial plexus block, with supraclavicular approach, which have reached the higher approval or given less side effects.

METHODS

The Author reviewed the following techniques: Kulenkampff's (1911), Winnie's (1970), Brown's (1988) and Alemanno's (1992).

CONCLUSIONS

The four methods that were presented in the last century differ not for the results, which are more or less the same, but for the possible side effects, which result to be less with the latest techniques.

[A new approach to the brachial plexus: clinical results]

The paper reports the results of a new brachial plexus blockade technique using a supraclavicular route which was carried out in 74 patients. From the analysis of results it can be seen that the technique is comparable to that of Moore and Winnie in terms of the success of the blockade. However, this technique appears to be safer in relation to complications such as pneumothorax and subarachnoid injection.

[A new approach to the brachial plexus]

In order to avoid complications such as pneumothorax or subarachnoid injection, the paper proposes a new approach to brachial plexus using a supraclavear route based on the pulse of the subclavian artery and 7th cervical vertebra. Complications are unlikely using this route.

[Cerebral edema: osmotic therapy in clinical practice]

Having described the therapeutic protocol to be followed in cases of cerebral edema, starting with aspecific measures (hyperventilation in a position of 30 degrees, etc.) for slight cases and progressing to specific therapy (osmotherapy, barbiturates) for more severe cases, the paper focuses on osmotherapy and illustrates the theoretical principles on which it is based and its use in clinical practice.