top of page

Prism Pillarグループ


Temporal Relation

Temporal relationships among depression, medical comorbidity, and death or cardiovascular disease (CVD) events are complex. Clarifying temporal relationships may enhance current insight regarding the nature of the association of depression with poor outcomes. The temporal relation of depression symptoms (DS; score

temporal relation

In the 2012 i2b2 Challenge, 310 discharge summaries were annotated for temporal information. The challenge focused specifically on the identification of clinically relevant events in the patient records, and the relative ordering of the events with respect to each other and with respect to time expressions included in the records. This task was broken down into two steps, each corresponding to a separate track: (1) extraction of events and time expressions and (2) identification of temporal relations. We also established an end to end track that combined both steps to evaluate state of the art in system performance in temporal information extraction. Eighteen teams participated in the temporal relations challenge (see online supplementary appendix table 1). The results of the challenge were presented in a workshop that i2b2 organized in co-sponsorship with the American Medical Informatics Association (AMIA), at the Fall Symposium of AMIA in 2012.

In the clinical domain, there have been some recent efforts to adapt TimeML annotation guidelines to clinical narratives.315 Although these corpora are mostly in pilot stage and size, they have proven the initial success in adopting TimeML style annotations to the clinical domain. In addition, researchers have also explored other alternatives to label temporal information in clinical text. Zhou et al2 proposed a temporal constraint structural representation that translates the temporal relations of events and time expressions in a narrative to temporal interval representation.16 Tao et al17 proposed a web ontology language, CNTRO, to describe temporal relations in clinical narratives. The consensus among these temporal representations315217 is that the following elements are the most critical to capture: clinically related events, time expressions with some value normalization, and the temporal relations between entities (events and time expressions). Thus for the 2012 i2b2 Challenge, we modified the TimeML guidelines to emphasize these three aspects: events, temporal expressions, and temporal relations.

Although the task of event detection and temporal relation classification for the general and clinical domains demand quite different methods, temporal expression (date, time, duration, and frequency) extraction and normalization in the clinical domain is not much different from that in the general domain, except for the medication dosage and frequency short hand widely used by clinical practitioners. In the general domain, the TempEval 2 shared tasks include a time expression detection track.13 The best performing temporal expression extraction system in TempEval 2 is HeidelTime18 which uses four sets of handcrafted rules to identify and classify temporal expressions.

For the TLINK classification task, F measure was used as the primary evaluation metric. Prior to evaluation, we compute the TC of the TLINKS provided by the system and the TC of the TLINKs found in the gold standard. The precision of the system output is the percentage of system TLINKs that can be verified in the TC of the gold standard TLINKs. Recall of the system output is the percentage of gold standard TLINKs that can be verified in the TC of the system TLINK output. We adapted the TLINK evaluation script for TempEval3 by UzZaman and Allen24 to compute temporal closure in our TLINK evaluation.

Analysis of the 25% sample records shows that the recognition of EVENT to section time TLINKs in general is easier than the recognition of other types of TLINKs (EVENT-EVENT, EVENT-TIMEX3, TIMEX3-TIMEX3, and TIMEX3-EVENT); 45.87% of the TLINKs in the sample gold standard records anchor EVENTs to section time. On average, 19.98 submissions out of the total 28 submissions correctly identified EVENT to section time TLINKs while only 11.92 out of 28 submissions correctly identified the other type of temporal relations. Among the non-section time TLINKs, EVENT-TIMEX3 and EVENT-EVENT relations were easier to detect, while TIMEX3-TIMEX3 and TIMEX3-EVENT were more challenging. Further analysis of the TLINKs suggests that this is likely due to the fact that many of the TIMEX3-TIMEX3 and TIMEX3-EVENT temporal relations involve the anchoring of relative dates and durations, a problem that is consistent with a similar issue in TIMEX3 extraction.

Objective:To create an end-to-end system to identify temporal relation in discharge summaries for the 2012 i2b2 challenge. The challenge includes event extraction, timex extraction and temporal relation identification.

Results:For event extraction, the system achieved 0.9415 (P), 0.8930 (R), 0.9166 (P&R) and 0.9166 (F), respectively. For timex extraction, it achieved 0.8818, 0.9489, 0.9141 and 0.9141, respectively. For temporal relation, it achieved 0.6589, 0.7129, 0.6767 and 0.6849, respectively. For end-to-end temporal relation, it achieved 0.5904, 0.5944, 0.5921 and 0.5924, respectively. As for F-measure as evaluation, we were ranked as the first out of 14 competing teams (event extraction), the first out of 14 teams (timex extraction), the third out of 12 teams (temporal relation), and the second out of 7 teams (end-to-end temporal relation).

Temporal relation extraction systems aim to identify and classify the temporal relation between a pair of entities provided in a text. For instance, in the sentence "Bob sent a message to Alice while she was leaving her birthday party." one can infer that the actions "sent" and "leaving" entails a temporal relation that can be described as "simultaneous".

The diagnosis of psychiatric disorders is currently based on a clinical and psychiatric examination (intake). Ancillary tests are used minimally or only to exclude other disorders. Here, we demonstrate a novel mathematical approach based on the field of p-adic numbers and using electroencephalograms (EEGs) to identify and differentiate patients with schizophrenia and depression from healthy controls. This novel approach examines spatio-temporal relations of single EEG electrode signals and characterizes the topological structure of these relations in the individual patient. Our results indicate that the relational topological structures, characterized by either the personal universal dendrographic hologram (DH) signature (PUDHS) or personal block DH signature (PBDHS), form a unique range for each group of patients, with impressive correspondence to the clinical condition. This newly developed approach results in an individual patient signature calculated from the spatio-temporal relations of EEG electrodes signals and might help the clinician with a new objective tool for the diagnosis of a multitude of psychiatric disorders.

The transition from theoretical modelling to practical applications is presented in an article by Shor et al.49. Here, the clustering algorithms and generated dendrograms thereof were used to represent hierarchic relations between events that consist of EEG measurements. The novel technique is based on a time series of dendrograms instead of straightforward use of an EEG-output time series. The medical diagnostic algorithm was based on a relatively rough dendrogram analysis known as quantum potential, which is a central concept of Bohmian mechanics. We interpreted the data according to Bohm and Hiley26. It should be noted that quantum probability and information are widely used in the modelling of cognition, decision-making, psychology and social sciences42,50,51,52,53 and are known as quantum-like to distinguish them from genuine quantum theory of cognition54,55. Bohr emphasized the possibility to apply quantum methodology in biology56 and Shor et al. described the quantum-like model as a dendrogramic configuration space20.

Using objective tools to classify psychiatric patients is a great challenge. We here present a novel algorithm using EEG to differentiate patients suffering from psychiatric disorders (depression and schizophrenia) and control participants. By means of a distance metric and a linkage algorithm as described in detail in the methods section, relations between events can be represented as tree structures called dendrograms.

It should be kept in mind that we currently analyze the relations between EEG-outputs (events) rather than the absolute magnitude of the outputs The relations are expressed by 2-adic numbers, each representing a complex context of spatial (locations of 19 electrodes) and temporal dynamics of the state of the brain. 041b061a72

  • グループについて


    bottom of page