Dataset Information

Breaking the circularity in circular analyses: Simulations and formal treatment of the flattened average approach.

ABSTRACT: There has been considerable debate and concern as to whether there is a replication crisis in the scientific literature. A likely cause of poor replication is the multiple comparisons problem. An important way in which this problem can manifest in the M/EEG context is through post hoc tailoring of analysis windows (a.k.a. regions-of-interest, ROIs) to landmarks in the collected data. Post hoc tailoring of ROIs is used because it allows researchers to adapt to inter-experiment variability and discover novel differences that fall outside of windows defined by prior precedent, thereby reducing Type II errors. However, this approach can dramatically inflate Type I error rates. One way to avoid this problem is to tailor windows according to a contrast that is orthogonal (strictly parametrically orthogonal) to the contrast being tested. A key approach of this kind is to identify windows on a fully flattened average. On the basis of simulations, this approach has been argued to be safe for post hoc tailoring of analysis windows under many conditions. Here, we present further simulations and mathematical proofs to show exactly why the Fully Flattened Average approach is unbiased, providing a formal grounding to the approach, clarifying the limits of its applicability and resolving published misconceptions about the method. We also provide a statistical power analysis, which shows that, in specific contexts, the fully flattened average approach provides higher statistical power than Fieldtrip cluster inference. This suggests that the Fully Flattened Average approach will enable researchers to identify more effects from their data without incurring an inflation of the false positive rate.

SUBMITTER: Bowman H

PROVIDER: S-EPMC7721178 | biostudies-literature | 2020 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Breaking the circularity in circular analyses: Simulations and formal treatment of the flattened average approach.

Bowman Howard H Brooks Joseph L JL Hajilou Omid O Zoumpoulaki Alexia A Litvak Vladimir V

PLoS computational biology 20201123 11

There has been considerable debate and concern as to whether there is a replication crisis in the scientific literature. A likely cause of poor replication is the multiple comparisons problem. An important way in which this problem can manifest in the M/EEG context is through post hoc tailoring of analysis windows (a.k.a. regions-of-interest, ROIs) to landmarks in the collected data. Post hoc tailoring of ROIs is used because it allows researchers to adapt to inter-experiment variability and dis ...[more]

PMID: 33226982

Similar Datasets

Project description:A part of current research has intensively been focused on the proliferation and metabolic processes governing biological systems. Since the advent of high throughput methodologies like microarrays, the load of genomic data has increased geometrically and along with that the need for computational methods which will interpret these data. In the present work we study in vitro the common proliferation and metabolic processes, which are combined to the common oncogenic pathways, as far as gene expression is concerned, between the T-cell acute lymphoblastic leukemia (CCRF-CEM) and the rhabdomyosarcoma (TE-671) cell lines. We present a computational approach, using cDNA microarrays, in order to identify commonalities between diverse biological systems. Our analysis predicted that JAK1, STAT1, PIAS2 and CDK4 are the driving forces in the two cell lines. This type of analysis can lead to the understanding of the common mechanisms that transform physiological cells to malignant, as well as it reveals a new holistic way to understandthe the dynamics of tumor onset as well as the mechanistics of oncogenic drivers. The present work is concerned with the common expressional profile of two cell lines: the T-cell acute lymphoblastic leukemia (CCRF-CEM) and the rhabdomyosarcoma (TE-671) cell lines. Our investigation was focused on the identification of genes that share a common expression profile between the two cell lines. Both cell lines are characterized by the fact that their differentiation has stopped at an early stage, before they mature to their final cell type. Normally, these cells would have matured and progressed into differentiated cells, constituting blood and muscle cells, respectively. At some unknown point, normal differentiation ceased for these cells and they became malignant. From that point on, to the first manifestation of symptoms of malignancy, there is a lack of knowledge regarding the mechanisms underlying oncogenesis. From these observations, the question whether two distinct cell types destined to fulfill different functions, manifest similar mechanisms of growth and progression due to their malignant character, arises. The present study was focused on the identification of the differential expression profiles underlying the two cell lines. A previous report studied the expression profile of seven ARMS cell lines possessing the PAX3-FKHR fusion gene, along with other cell lines of different tumor types (22). To our knowledge, this is the first time that a comparison between two totally different types of neoplasia, such as the CCRF-CEM and TE-671 cell lines, is attempted. These mechanisms are examined with purpose to identify common drivers that lead to the progression of tumor cells. We hereby propose a new computational approach for the investigation of common oncogenic drivers.

Project description:BackgroundAlthough "buffering" as a homeostatic mechanism is a universal phenomenon, the quantitation of buffering action remains controversial and problematic. Major shortcomings are: lack of a buffering strength unit for some buffering phenomena, multiple and mutually incommensurable units for others, and lack of a genuine ratio scale for buffering strength. Here, I present a concept of buffering that overcomes these shortcomings.TheoryBriefly, when, for instance, some "free" H+ ions are added to a solution (e.g. in the form of strong acid), buffering is said to be present when not all H+ ions remain "free" (i.e., bound to H2O), but some become "bound" (i.e., bound to molecules other than H2O). The greater the number of H+ ions that become "bound" in this process, the greater the buffering action. This number can be expressed in two ways: 1) With respect to the number of total free ions added as "buffering coefficient b", defined in differential form as b = d(bound)/d(total). This measure expresses buffering action from nil to complete by a dimensionless number between 0 and 1, analogous to probabilites. 2) With respect to the complementary number of added ions that remain free as "buffering ratio B", defined as the differential B = d(bound)/d(free). The buffering ratio B provides an absolute ratio scale, where buffering action from nil to perfect corresponds to dimensionless numbers between 0 and infinity, and where equal differences of buffering action result in equal intervals on the scale. Formulated in purely mathematical, axiomatic form, the concept reveals striking overlap with the mathematical concept of probability. However, the concept also allows one to devise simple physical models capable of visualizing buffered systems and their behavior in an exact yet intuitive way.ConclusionThese two measures of buffering action can be generalized easily to any arbitrary quantity that partitions into two compartments or states, and are thus suited to serve as standard units for buffering action. Some exemplary treatments of classical and non-classical buffering phenomena are presented in the accompanying paper.

Dataset Information

Breaking the circularity in circular analyses: Simulations and formal treatment of the flattened average approach.

Publications

Breaking the circularity in circular analyses: Simulations and formal treatment of the flattened average approach.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets