Dissertation Defense


Ali Hummos






2206C Student Center


The hippocampus, an area in the temporal lobe of the mammalian brain, participates in the storage of personal memories and life events, including traumatic memories and the consequent symptoms of post-traumatic stress, giving importance to the study of the machinery of hippocampal memory storage and retrieval. The circuit is known to be controlled by the neuromodulator Acetylcholine, which switches the circuit between the memory storage state and the memory retrieval state.

We built a computational model of the hippocampus with the ability to perform both memory storage and retrieval functions, controlled by the level of Acetylcholine. This functional separation decrease interference between the two circuit functions while sharing the same physical implementation of a network of spiking neurons.

We discovered three important differences between the storage and retrieval circuits. First, they had difference in how they produced runaway excitation, an aberrant spread of brain activity leading to seizures. Second, the two circuits had distinct mechanisms to maintain control over runaway excitation spread. These two findings provided the first classification of seizures based on the functional state of the brain, and suggested the need for specific treatments for each type.

Third, we found the two circuits also had unique ways of generating theta rhythmic activity, which is theorized to have a fundamental role in memory storage and retrieval. Our model uncovered an unexpected complexity in theta rhythm generation across functional states of the circuit. These findings can allow for deciphering the computations carried out by the circuit, based on the engaged mechanisms of rhythm generation.

Seminar Series


Pericles Giannaris






2206A Student Center

Use of the N-ary Relational Schema to Atomize Compound Relational Triples

Electronic medical records document health information in structured format and in unstructured free text format.  Health information in structured format contains laboratory results, vital signs, patient demographics etc.  The unstructured free text is the prime source of healthcare information documenting providers’ interpretations of health conditions, diagnoses, medical interventions, impressions, etc.  In order to uncover unknown information and search for patterns in health data with computational methods, we need to structure the unstructured free text data.  For that, we use information extraction, a computational technique for analyzing free text and deriving structured information.  Extracted information from free text can be represented in the form of relational triples.  Relational triples are statements of a single fact composed of subject-relation-object.  These triple statements allow the development of knowledge bases, knowledge graphs or the application of inference rules.  In our research, we employ Stanford’s CoreNLP engine for information extraction in triple format.  This format helps us to develop Resource Description Framework (RDF) networks where each subject and object become nodes and the edges represent the relations between the nodes. However, most of the triples produced by CoreNLP convey multiple facts (compound triple), instead of a single fact (atomic triple). Compound triples produce networks with nodes representing multiple entities instead of a single entity causing issues of network representation of our data. Here, we extend the use of CoreNLP to atomize compound triples.  Our approach is based on the N-ary relational schema that links an individual to multiple individuals or values. Our approach includes triple decomposition and ontological modeling.