The fields of Computer Science and Neural Science have a symbiotic relationship. Many strategies used in computing systems, especially in artificial intelligence, are based in biological principles: neuroprocessors made from neural cultures, neuroprosthetics based on ganglion cell recordings, robotic guiding based on biological neural maps, etc. On the other hand, we can find these AI computing strategies applied to biological and medical problems; for example detecting neural preliminary pathologies using speech traces, EEG signals, or neuroimaging techniques. This Research Topic will focus on intelligent systems based on neural evidences or intelligent techniques for detecting neural pathologies. This synergistic approach will permit getting inside the main mechanism and principles of neural systems.
Machine learning is a type of this advanced computational algorithm. Neuron-like elements could learn to achieve meaningful computations and to extract useful features from images or sound waves. Three different machine learning problems can be identified: the inference problem with prediction objectives, the parameter-learning (maybe by changing synapse strengths) problem, or the structure-learning problem, made by evolving systems for determining the structure.
Machine learning has two very different relationships to neuroscience. As with any other science, modern machine-learning methods can be very helpful in analyzing the data or predicting neurological degeneration. Some examples of this include sorting the spikes picked up by an extracellular electrode into spike trains from different neurons, or estimating the probability of Parkinson’s or Alzheimer’s disease based on MRI, EEG or speech signals. A much more interesting relationship is the use of machine-learning algorithms as a source of theories about how the brain works, embedding neural models in a robot behavior, artificial vision systems or affective computation platforms.
Recently, new machine learning techniques proposed several unsupervised methods for creating multiple layers of features without requiring any labels. These methods are significantly better than traditional methods for creating high-level features, so 'deep' learning is making a comeback for tasks like the inference problem, the parameter problem or the structure-learning problem. These systems significantly enrich the interaction between machine learning and neuroscience.
We welcome contributions on, but not limited to, the following subjects:
- Machine learning for predicting neurological disease based on MRI, EEG or Speech
- Embedding neural models in artificial or biological systems.
- Deep Learning for neural data
- New tools for Brain-Computer Interfaces
The papers should be in the form of original research articles, review articles, technology reports, perspectives, case reports, also in a short form.
The fields of Computer Science and Neural Science have a symbiotic relationship. Many strategies used in computing systems, especially in artificial intelligence, are based in biological principles: neuroprocessors made from neural cultures, neuroprosthetics based on ganglion cell recordings, robotic guiding based on biological neural maps, etc. On the other hand, we can find these AI computing strategies applied to biological and medical problems; for example detecting neural preliminary pathologies using speech traces, EEG signals, or neuroimaging techniques. This Research Topic will focus on intelligent systems based on neural evidences or intelligent techniques for detecting neural pathologies. This synergistic approach will permit getting inside the main mechanism and principles of neural systems.
Machine learning is a type of this advanced computational algorithm. Neuron-like elements could learn to achieve meaningful computations and to extract useful features from images or sound waves. Three different machine learning problems can be identified: the inference problem with prediction objectives, the parameter-learning (maybe by changing synapse strengths) problem, or the structure-learning problem, made by evolving systems for determining the structure.
Machine learning has two very different relationships to neuroscience. As with any other science, modern machine-learning methods can be very helpful in analyzing the data or predicting neurological degeneration. Some examples of this include sorting the spikes picked up by an extracellular electrode into spike trains from different neurons, or estimating the probability of Parkinson’s or Alzheimer’s disease based on MRI, EEG or speech signals. A much more interesting relationship is the use of machine-learning algorithms as a source of theories about how the brain works, embedding neural models in a robot behavior, artificial vision systems or affective computation platforms.
Recently, new machine learning techniques proposed several unsupervised methods for creating multiple layers of features without requiring any labels. These methods are significantly better than traditional methods for creating high-level features, so 'deep' learning is making a comeback for tasks like the inference problem, the parameter problem or the structure-learning problem. These systems significantly enrich the interaction between machine learning and neuroscience.
We welcome contributions on, but not limited to, the following subjects:
- Machine learning for predicting neurological disease based on MRI, EEG or Speech
- Embedding neural models in artificial or biological systems.
- Deep Learning for neural data
- New tools for Brain-Computer Interfaces
The papers should be in the form of original research articles, review articles, technology reports, perspectives, case reports, also in a short form.
