Johan Himberg

Validating the independent components of neuroimaging time series via clustering and visualization.

Recently, independent component analysis (ICA) has been widely used in the analysis of brain imaging data. An important problem with most ICA algorithms is, however, that they are stochastic; that is, their results may be somewhat different in different runs of the algorithm. Thus, the outputs of a single run of an ICA algorithm should be interpreted with some reserve, and further analysis of the algorithmic reliability of the components is needed. Moreover, as with any statistical method, the results are affected by the random sampling of the data, and some analysis of the statistical significance or reliability should be done as well. Here we present a method for assessing both the algorithmic and statistical reliability of estimated independent components. The method is based on running the ICA algorithm many times with slightly different conditions and visualizing the clustering structure of the obtained components in the signal space. In experiments with magnetoencephalographic (MEG) and functional magnetic resonance imaging (fMRI) data, the method was able to show that expected components are reliable; furthermore, it pointed out components whose interpretation was not obvious but whose reliability should incite the experimenter to investigate the underlying technical or physical phenomena. The method is implemented in a software package called Icasso.

Independent component analysis of fMRI group studies by self-organizing clustering

Independent component analysis (ICA) is a valuable technique for the multivariate data-driven analysis of functional magnetic resonance imaging (fMRI) data sets. Applications of ICA have been developed mainly for single subject studies, although different solutions for group studies have been proposed. These approaches combine data sets from multiple subjects into a single aggregate data set before ICA estimation and, thus, require some additional assumptions about the separability across subjects of group independent components. Here, we exploit the application of similarity measures and a related visual tool to study the natural self-organizing clustering of many independent components from multiple individual data sets in the subject space. Our proposed framework flexibly enables multiple criteria for the generation of group independent components and their random-effects evaluation. We present real visual activation fMRI data from two experiments, with different spatiotemporal structures, and demonstrate the validity of this framework for a blind extraction and selection of meaningful activity and functional connectivity group patterns. Our approach is either alternative or complementary to the group ICA of aggregated data sets in that it exploits commonalities across multiple subject-specific patterns, while addressing as much as possible of the intersubject variability of the measured responses. This property is particularly of interest for a blind group and subgroup pattern extraction and selection.

Recognizing human motion with multiple acceleration sensors

In this paper experiments with acceleration sensors are described for human activity recognition of a wearable device user. The use of principal component analysis and independent component analysis with a wavelet transform is tested for feature generation. Recognition of human activity is examined with a multilayer perceptron classifier. Best classification results for recognition of different human motion were 83-90%, and they were achieved by utilizing independent component analysis and principal component analysis. The difference between these methods turned out to be negligible.

Time series segmentation for context recognition in mobile devices

Recognizing the context of use is important in making mobile devices as simple to use as possible. Finding out what the users situation is can help the device and underlying service in providing an adaptive and personalized user interface. The device can infer parts of the context of the user from sensor data: the mobile device can include sensors for acceleration, noise level, luminosity, humidity, etc. In this paper we consider context recognition by unsupervised segmentation of time series produced by sensors. Dynamic programming can be used to find segments that minimize the intra-segment variances. While this method produces optimal solutions, it is too slow for long sequences of data. We present and analyze randomized variations of the algorithm. One of them, global iterative replacement or GIR, gives approximately optimal results in a fraction of the time required by dynamic programming. We demonstrate the use of time series segmentation in context recognition for mobile phone applications.

Icasso: software for investigating the reliability of ICA estimates by clustering and visualization

A major problem in application of independent component analysis (ICA) is that the reliability of the estimated independent components is not known. Firstly, the finite sample size induces statistical errors in the estimation. Secondly, as real data never exactly follows the ICA model, the contrast function used in the estimation may have many local minima which are all equally good, or the practical algorithm may not always perform properly, for example getting stuck in local minima with strongly suboptimal values of the contrast function. We present an explorative visualization method for investigating the relations between estimates from FastICA. The algorithmic and statistical reliability is investigated by running the algorithm many times with different initial values or with differently bootstrapped data sets, respectively. Resulting estimates are compared by visualizing their clustering according to a suitable similarity measure. Reliable estimates correspond to tight clusters, and unreliable ones to points which do not belong to any such cluster. We have developed a software package called Icasso to implement these operations. We also present results of this method when applying Icasso on biomedical data.

A SOM based cluster visualization and its application for false coloring

The self-organizing map (SOM) is widely used as a data visualization method in various engineering applications. It performs a nonlinear mapping from a high-dimensional data space to a lower dimensional visualization space. In this paper, a simple method for visualizing the cluster structure of SOM model vectors is presented. The method may be used to produce tree-like visualizations, but the main application here is to derive different color coding that express the approximate cluster structure of the SOM model vectors. This coloring may be exploited in making false color (pseudo color) presentations of the original data. The method is especially designed as an easily implementable, explorative cluster visualization tool.

On-line personalization of a touch screen based keyboard

The user expectations for usability and personalization along with decreasing size of handheld devices challenge traditional keypad layout design. We have developed a method for on-line adaptation of a touch pad keyboard layout. The method starts from an original layout and monitors the usage of the keyboard by recording and analyzing the keystrokes. An on-line learning algorithm subtly moves the keys according to the spatial distribution of keystrokes. In consequence, the keyboard matches better to the users physical extensions and grasp of the device, and makes the physical trajectories during typing more comfortable. We present two implementations that apply different vector quantization algorithms to produce an adaptive keyboard with visual on-line feedback. Both qualitative and quantitative results show that the changes in the keyboard are consistent, and related to the users handedness and hand extensions. The testees found the on-line personalization positive. The method can either be applied for on-line personalization of keyboards or for ergonomics research

Unsupervised clustering of symbol strings and context recognition

The representation of information based on symbol strings has been applied to the recognition of context. A framework for approaching the context recognition problem has been described and interpreted in terms of symbol string recognition. The symbol string clustering map (SCM) is introduced as an efficient algorithm for the unsupervised clustering and recognition of symbol string data. The SCM can be implemented in an online manner using a computationally simple similarity measure based on a weighted average. It is shown how measured sensor data can be processed by the SCM algorithm to learn, represent and distinguish different user contexts without any user input.

Collaborative context determination to support mobile terminal applications

Mobile devices, together with their users, are constantly moving from one situation to another. To adapt applications to these changing contexts, the devices must have ways to recognize the contexts. There are various sources for context information: sensors, tags, positioning systems, to name a few. The raw signals from these sources are translated into higher-level interpretations of the situation. Unfortunately, such data is often unreliable and constantly changing. We seek to improve the reliability of context recognition through an analogy to human behavior. Where multiple devices are around, they can jointly negotiate on a suitable context and behave accordingly. This approach is becoming particularly attractive with the multitude of personal devices on the market. We present a collaborative context determination scheme, suggest examples of potential applications of such collaborative behavior, and raise issues of context recognition, context communication, and network requirements.

Collaborative context recognition for handheld devices

Handheld communication devices equipped with sensing capabilities can recognize some aspects of their context to enable novel applications. We seek to improve the reliability of context recognition through an analogy to human behavior. Where multiple devices are around, they can jointly negotiate on a suitable context and behave accordingly. We have developed a method for this collaborative context recognition for handheld devices. The method determines the need to request and collaboratively recognize the current context of a group of handheld devices. It uses both local context time history information and spatial context information of handheld devices within a certain area. The method exploits dynamic weight parameters that describe content and reliability of context information. The performance of the method is analyzed using artificial and real context data. The results suggest that the method is capable of improving the reliability.