Alex Frid

Fusing enacted and expected mimicry generates a winning strategy that promotes the evolution of cooperation

Although cooperation and trust are essential features for the development of prosperous populations, they also put cooperating individuals at risk for exploitation and abuse. Empirical and theoretical evidence suggests that the solution to the problem resides in the practice of mimicry and imitation, the expectation of opponent’s mimicry and the reliance on similarity indices. Here we fuse the principles of enacted and expected mimicry and condition their application on two similarity indices to produce a model of mimicry and relative similarity. Testing the model in computer simulations of behavioral niches, populated with agents that enact various strategies and learning algorithms, shows how mimicry and relative similarity outperforms all the opponent strategies it was tested against, pushes noncooperative opponents toward extinction, and promotes the development of cooperative populations. The proposed model sheds light on the evolution of cooperation and provides a blueprint for intentional induction of cooperation within and among populations. It is suggested that reducing conflict intensities among human populations necessitates (i) instigation of social initiatives that increase the perception of similarity among opponents and (ii) efficient lowering of the similarity threshold of the interaction, the minimal level of similarity that makes cooperation advisable.

An SVM based algorithm for analysis and discrimination of dyslexic readers from regular readers using ERPs

Dyslexia is a learning disability that impairs a persons ability to decode words accurately and fluently. This deficit can manifest itself in the language-related domain as difficulties in phonological and orthographic working memory, brain systems asynchrony, poor executive function skills and/or poor rapid naming processing. However it is not clear yet whether the dyslexia phenomenon is only related to language or if it can also be seen as a non-language deficit. Moreover, if it is also related to non-language activity, it is important to verify if it is possible to identify dyslexic readers at the earliest stage of information processing for better and effective remediation. Based on this, an effective algorithm was developed for analysis and classification of subjects as either Regular Readers or Dyslexic Readers, by using EEG recorded channels with Event Related Potentials (ERP) methodology during an auditory, short non-linguistic, simple, sub-phonetic choices reaction time task.

Temporal pattern recognition via temporal networks of temporal neurons

We show that real valued continuous functions can be recognized in a reliable way, with good generalization ability using an adapted version of the Liquid State Machine (LSM) that receives direct real valued input. Furthermore this system works without the necessity of preliminary extraction of signal processing features. This avoids the necessity of discretization and encoding that has plagued earlier attempts on this process. We show this is effective on a simulated signal designed to have the properties of a physical trace of human speech. The main changes to the basic liquid state machine paradigm are (i) external stimulation to neurons by normalized real values and (ii) adaptation of the integrate and fire neurons in the liquid to have a history dependent sliding threshold (iii) topological constraints on the network connectivity.

Diagnosis of Parkinson's disease from continuous speech using deep convolutional networks without manual selection of features

Parkinsons Disease (PD) is a relatively common neurodegenerative disabling disease. It affects central nervous system with profound effect on the motor system. The most common symptoms include slowness, rigidity and tremor during motion. It has been suggested that the vocal cords are among the first one to be affected and thus the speech is affected at very early stage of the disease and continues to deteriorate as the disease progress. In this work, we focus on automating the process of diagnosis from continuous native speech by removing the necessity of a trained personal from the diagnosis process. This is done by using an adaptation of Convolutional Neural Network (CNN) architecture for one-dimensional signal processing (i.e. raw speech signal) on a relatively small training set. This is a continuation to previous works where we showed (i) that this task can be achieved by using manually-extracted features of the speech (such as formants and their ratios) and (ii) by using an automatic process of auditory features extraction, where the features were selected by signal processing specialist.

Classification from generation: Recognizing deep grammatical information during reading from rapid event-related fMRI.

A novel fMRI classification method designed for rapid event related fMRI experiments is described and applied to the classification of loud reading of isolated words in Hebrew. Three comparisons of different grammatical complexity were performed: (i) words versus asterisks (ii) “with diacritics versus without diacritics” and (iii) “with root versus no root”. We discuss the most difficult task and, for comparison, the easiest one. Earlier work using more standard classification techniques (machine learning and statistical) succeeded fully only in the simplest of these tasks (i), but produced only partial results on (ii) and failed completely, even on the training set on the deepest task (iii). The method performs a “best match” between pre-processed data and computing a full library of artificially generated examples. The method involves a deconvolution of the rapid events on the data and performing a convolution on the generated data. The best-match is performed over all “words” constructed by convolving the response functions of each value of each event performed in a “windowed” sequence. This is accomplished separately for all voxels and then a voting procedure defines the outcome. Using the same feature selection (ANOVA) as in the earlier methods, (i) there is a dramatic increase in the accuracy rate for the third (most difficult task) on the intra-run level (88%) as well as the first task (ii) Unlike the earlier methods training and testing over all runs (within subject) achieves a significant level of classification (64% accuracy) for the training set. This shows the information for this “deeper” cognitive task can in fact be extracted from the fMRI information.

Recognizing deep grammatical information during reading from event related fMRI

This experiment was designed to see if information related to linguistic characteristics of read text can be deduced from fMRI data via machine learning techniques. Individuals were scanned while reading text the size of words in loud reading. Three experiments were performed corresponding to different degrees of grammatical complexity that is performed during loud reading: (1) words and pseudo-words were presented to subjects; (2) words with diacritical marking and words without diacritical markings were presented to subjects; (3) Hebrew words with Hebrew root and Hebrew words without Hebrew root were presented to subjects. The working hypothesis was that the more complex the needed grammatical processing needed, the more difficult it should be to perform the classification at the level of temporal and spatial resolution given by an fMRI signal. We were able to accomplish the first task completely. The second and third task did not succeed when all the data is used simultaneously. However, the third task was successful when training and testing was done within a continuous scanning run. (The experimental protocol did not allow this for the second task.) This does establish that complex linguistic information is decodable from fMRI scans. On the other hand, the need to restrict to the intra-run situation indicates that additional work is needed to compensate for distortions introduced between scanning runs.

Machine learning techniques and the existence of variant processes in humans declarative memory

This work uses supervised machine learning methods over fMRI brain scans to establish the existence of two different encoding procedures for human declarative memory. Declarative knowledge refers to the memory for facts and events and initially depends on the hippocampus. Recent studies which used patients with hippocampal lesions and neuroimaging data, suggested the existence of an alternative process to form declarative memories. This process is triggered by learning mechanism called “Fast Mapping (FM)”, as opposed to the ‘standard’ “Explicit Encoding (EE)” learning procedure. The present work gives a clear biomarker on the existence of two distinct encoding procedures as we can accurately predict which of the processes is being used directly from voxel activity in fMRI scans. The scans are taken during retrieval of information wherein the tasks are identical regardless of which procedure was used for acquisition and by that reflect conclusive prediction. This is an identification of a more subtle cognitive task than direct perceptual cognitive tasks as it requires some encoding and processing in the brain.

Differences in phase synchrony of brain regions between regular and dyslexic readers

Dyslexia is a learning disability that impairs a persons ability to decode words accurately and fluently. This deficit can manifest itself as difficulties in orthographic working memory, brain systems asynchrony, poor executive function skills and/or rapid naming processing. Although each of the aforementioned factors incorporates different brain systems and is activated at different speeds and in different manners, it is clear that for good reading performance to occur, these systems must be activated in a synchronized manner (hence, the name “A-synchronization Theory”). This aforementioned theory of A-synchronization indicates that impairment among dyslexic readers involves not only the speed of processing but also the integrating and processing of information emerging from different sensory systems, specifically, the visual and auditory modalities. Based on this, an effective algorithm was developed for analysis and classification of subjects as either Regular Readers or Dyslexic Readers. This was achieved by measuring phase synchrony in Electroencephalographic (EEG) recorded channels during visual Lexical Decision task. An average correct identification rate of 89% ± 2% was achieved when phase synchrony features from the gamma-band (30-40Hz) were extracted. This may indicate a lack of synchrony between different modalities in dyslexic readers related to attention process.