Chromosomal imbalances detected in NTRK-rearranged sarcomas by the use of comparative genomic hybridisation.

AIMS: NTRK-rearranged sarcomas are emerging as a distinct class of sarcomas of particular importance in the era of targeted therapy. The aim of this study was to use array comparative genomic hybridisation (aCGH) to explore the cytogenetic profile of six adult soft tissue sarcomas harbouring NTRK gene fusions. METHODS AND RESULTS: aCGH was performed on six adult soft tissue sarcomas with proven NTRK rearrangements [NTRK1, n = 1 (TPM3-NTRK1); NTRK2, n = 1 (MTMR2-NTRK2); NTRK3, n = 4 (two ETV6-NTRK3; two with unknown partners). The morphological patterns of these cases included inflammatory myofibroblastic tumour-like, fibrosarcoma/malignant peripheral nerve sheath tumour-like, and Ewing sarcoma-like. On the basis of the number of chromosomal copy number variations (CNVs), ranging from two to 15 per sample, NTRK-associated sarcomas could be subdivided into two groups: one with a relatively simple karyotype (n = 2; median of three genomic alterations), and those with a more complex karyotype (n = 4; median of 11 genomic imbalances). Recurrent chromosomal CNVs included gains at chromosomes 6p, 1q, 7 (whole chromosome), and 12p, and losses at chromosomes 10q, 13q, 19q, and 9p. CONCLUSIONS: NTRK-rearranged sarcomas constitute a heterogeneous group of tumours that can show a relatively simple or a complex karyotype. Although there were some, but inconsistent, associations between karyotype complexity and morphology, our study showed that a more complex karyotype in this group of tumours appeared to correlate with more aggressive clinical behaviour. Gains at chromosome 6p and 1q were the most common recurrent genomic alterations, being present in 67% of the samples (4/6), followed by gains at chromosome 7, which were present in 50% of the samples (3/6).

DNA damage-inducible transcript 3 immunohistochemistry is highly sensitive for the diagnosis of myxoid liposarcoma but care is required in interpreting the significance of focal expression.

AIMS: Myxoid liposarcoma (MLPS) is characterised by DNA damage-inducible transcript 3 (DDIT3) gene rearrangements, confirmation of which is commonly used diagnostically. Recently, DDIT3 immunohistochemistry (IHC) has been reported to be highly sensitive and, when strict criteria are employed, specific for the diagnosis of MLPS. The aim of this study was to independently investigate DDIT3 IHC as a diagnostic marker for MLPS. METHODS AND RESULTS: DDIT3 IHC was performed on 52 MLPS and on 152 mimics on whole sections, and on 515 non-MLPS sarcomas in tissue microarray format. Only one MLPS (which had undergone acid-based decalcification) was completely negative. With inclusion of this case if any nuclear expression is considered to indicate positivity, the overall sensitivity of DDIT3 is 98% (51 of 52 cases) and the specificity is 94% (633 of 667 non-MLPS cases are negative). If a cut-off of >10% of neoplastic cells is required for positivity, then the sensitivity remains 98% (51/52) and the specificity is 98.5% (657 of 667 non-MLPS cases are negative). If a cut-off of >50% of cells is required for positivity, then the sensitivity is 96% (50 of 52 cases) but the specificity improves to 100%. CONCLUSIONS: Diffuse nuclear DDIT3 expression occurs in the overwhelming majority of MLPSs, and can be used to confirm the diagnosis in most cases without the need for molecular testing. A complete absence of expression argues strongly against MLPS, and almost completely excludes this diagnosis, particularly if there is consideration of technical factors such as decalcification. The significance of focal DDIT3 expression should be interpreted in the morphological and clinical context, although most tumours showing only focal expression are not MLPS.

Brain-informed speech separation (BISS) for enhancement of target speaker in multitalker speech perception.

Hearing-impaired people often struggle to follow the speech stream of an individual talker in noisy environments. Recent studies show that the brain tracks attended speech and that the attended talker can be decoded from neural data on a single-trial level. This raises the possibility of "neuro-steered" hearing devices in which the brain-decoded intention of a hearing-impaired listener is used to enhance the voice of the attended speaker from a speech separation front-end. So far, methods that use this paradigm have focused on optimizing the brain decoding and the acoustic speech separation independently. In this work, we propose a novel framework called brain-informed speech separation (BISS)1 in which the information about the attended speech, as decoded from the subject's brain, is directly used to perform speech separation in the front-end. We present a deep learning model that uses neural data to extract the clean audio signal that a listener is attending to from a multi-talker speech mixture. We show that the framework can be applied successfully to the decoded output from either invasive intracranial electroencephalography (iEEG) or non-invasive electroencephalography (EEG) recordings from hearing-impaired subjects. It also results in improved speech separation, even in scenes with background noise. The generalization capability of the system renders it a perfect candidate for neuro-steered hearing-assistive devices.

Multiway canonical correlation analysis of brain data.

Brain data recorded with electroencephalography (EEG), magnetoencephalography (MEG) and related techniques often have poor signal-to-noise ratios due to the presence of multiple competing sources and artifacts. A common remedy is to average responses over repeats of the same stimulus, but this is not applicable for temporally extended stimuli that are presented only once (speech, music, movies, natural sound). An alternative is to average responses over multiple subjects that were presented with identical stimuli, but differences in geometry of brain sources and sensors reduce the effectiveness of this solution. Multiway canonical correlation analysis (MCCA) brings a solution to this problem by allowing data from multiple subjects to be fused in such a way as to extract components common to all. This paper reviews the method, offers application examples that illustrate its effectiveness, and outlines the caveats and risks entailed by the method.

Decoding the auditory brain with canonical component analysis.

The relation between a stimulus and the evoked brain response can shed light on perceptual processes within the brain. Signals derived from this relation can also be harnessed to control external devices for Brain Computer Interface (BCI) applications. While the classic event-related potential (ERP) is appropriate for isolated stimuli, more sophisticated "decoding" strategies are needed to address continuous stimuli such as speech, music or environmental sounds. Here we describe an approach based on Canonical Correlation Analysis (CCA) that finds the optimal transform to apply to both the stimulus and the response to reveal correlations between the two. Compared to prior methods based on forward or backward models for stimulus-response mapping, CCA finds significantly higher correlation scores, thus providing increased sensitivity to relatively small effects, and supports classifier schemes that yield higher classification scores. CCA strips the brain response of variance unrelated to the stimulus, and the stimulus representation of variance that does not affect the response, and thus improves observations of the relation between stimulus and response.

Cortical hemispheric asymmetries are present at young ages and further develop into adolescence.

Specialization of the auditory cortices for pure tone listening may develop with age. In adults, the right hemisphere dominates when listening to pure tones and music; we thus hypothesized that (a) asymmetric function between auditory cortices increases with age and (b) this development is specific to tonal rather than broadband/non-tonal stimuli. Cortical responses to tone-bursts and broadband click-trains were recorded by multichannel electroencephalography in young children (5.1 ± 0.8 years old) and adolescents (15.2 ± 1.7 years old) with normal hearing. Peak dipole moments indicating activity strength in right and left auditory cortices were calculated using the Time Restricted, Artefact and Coherence source Suppression (TRACS) beamformer. Monaural click-trains and tone-bursts in young children evoked a dominant response in the contralateral right cortex by left ear stimulation and, similarly, a contralateral left cortex response to click-trains in the right ear. Responses to tone-bursts in the right ear were more bilateral. In adolescents, peak activity dominated in the right cortex in most conditions (tone-bursts from either ear and to clicks from the left ear). Bilateral activity was evoked by right ear click stimulation. Thus, right hemispheric specialization for monaural tonal stimuli begins in children as young as 5 years of age and becomes more prominent by adolescence. These changes were marked by consistent dipole moments in the right auditory cortex with age in contrast to decreases in dipole activity in all other stimulus conditions. Together, the findings reveal increasingly asymmetric function for the two auditory cortices, potentially to support greater cortical specialization with development into adolescence.

Bilateral input protects the cortex from unilaterally-driven reorganization in children who are deaf.

Unilateral hearing in childhood restricts input along the bilateral auditory pathways, possibly causing permanent reorganization. In this study we asked: (i) do the auditory pathways develop abnormally in children who are bilaterally deaf and hear with a unilateral cochlear implant? and (ii) can such differences be reversed by restoring input to the deprived ear? We measured multichannel electroencephalography in 34 children using cochlear implants and seven normal hearing peers. Dipole moments of activity became abnormally high in the auditory cortex contralateral to the first implant as unilateral cochlear implant use exceeded 1.5 years. This resulted in increased lateralization of activity to the auditory cortex contralateral to the stimulated ear and a decline in normal contralateral activity in response to stimulation from the newly implanted ear, corresponding to poorer speech perception. These results reflect an abnormal strengthening of pathways from the stimulated ear in consequence to the loss of contralateral activity including inhibitory processes normally involved in bilateral hearing. Although this reorganization occurred within a fairly short period (∼1.5 years of unilateral hearing), it was not reversed by long-term (3-4 years) bilateral cochlear implant stimulation. In bilateral listeners, effects of side of stimulation were assessed; children with long periods of unilateral cochlear implant use prior to bilateral implantation showed a reduction in normal dominance of contralateral input in the auditory cortex ipsilateral to the stimulated ear, further confirming an abnormal strengthening of pathways from the stimulated ear. By contrast, cortical activity in children using bilateral cochlear implants after limited or no unilateral cochlear implant exposure normally lateralized to the hemisphere contralateral to side of stimulation and retained normal contralateral dominance of auditory input in both hemispheres. Results demonstrate that the immature human auditory cortex reorganizes, potentially permanently, with unilateral stimulation and that bilateral auditory input provided with limited delay can protect the brain from such changes. These results indicate for the first time that there is a sensitive period for bilateral auditory input in human development with implications for functional hearing.

Expression of Melan-A in cutaneous granular cell tumours: a diagnostic pitfall.

Morphological overlap exists between cutaneous granular cell tumours (GCT) and malignant melanoma, with the melanocyte-specific markers HMB45 and Melan-A commonly used to support the diagnosis of melanoma. We recently encountered several cases of GCT in our practice showing strong expression of Melan-A. The aim of this study was to establish the prevalence of positive immunohistochemical staining for Melan-A and HMB45 in a series of unequivocal GCTs. We also aimed to assess the prevalence of staining for PRAME (PReferentially expressed Antigen in MElanoma), a marker expressed in >80% of primary melanomas as well as many non-melanocytic tumours. A total of 20 cutaneous/subcutaneous GCTs were evaluated using Melan-A, HMB45 and PRAME immunohistochemistry. Staining for Melan-A and HMB45 was scored using a semiquantitative scale from 0 (absent) to 3+ (staining present in >50% of tumour cells). PRAME expression was recorded as either positive (>75% of cell nuclei staining) or negative. Melan-A expression was observed in four GCTs (20%), with strong and diffuse (3+) staining seen in two cases (10%), both from anogenital areas. Weak patchy nuclear PRAME expression was seen in every case, interpreted to be negative. HMB45 was also negative in all cases (100%). Our study demonstrates that Melan-A expression can be strong and diffuse in a subset of otherwise unequivocal cutaneous GCTs, which may cause diagnostic confusion with malignant melanoma. HMB45 and PRAME did not stain any of the GCTs in our series.

Kinase expression in angiomatoid fibrous histiocytoma: panTRK is commonly expressed in the absence of NTRK rearrangement.

Angiomatoid fibrous histiocytoma (AFH) is a soft tissue tumour of intermediate (rarely metastasising) malignant potential, which harbours EWSR1/FUS gene fusions. These tumours can express anaplastic lymphoma kinase (ALK) in the absence of gene rearrangement or copy number alteration and can also coexpresses Pan-TRK immunohistochemistry (IHC). All EWSR1/FUS-rearranged AFH were retrieved from the files of three institutions and Pan-TRK (EPR17341), ALK and BRAF V600E IHC were performed. Fourteen AFH cases were identified, which included three cases of intracranial mesenchymal tumours with FET-CREB fusions. PanTRK and ALK positive immunostaining was identified in 9 (64.2%) and 12 (85.7%) cases, respectively. No NTRK or ALK translocations or increased copy number/amplification were identified in all eight cases which had fluorescence in situ hybridisation and/or next generation sequencing for NTRK1-3 and ALK available for assessment. None of the cases expressed BRAF-V600E.Although our study is limited, our report is the first to document PanTRK expression in AFH in the absence of identifiable NTRK1-3 gene alterations.

Keratin-Positive Giant Cell Tumor of Bone and Soft Tissue With HMGA2::NCOR2 Fusion in Children Under 10 With Response to Imatinib Therapy: A Case Series.

HMGA2::NCOR2 keratin-positive giant cell tumors in children with response to imatinib in an infant.

Crohn's disease activity: quantitative contrast-enhanced ultrasound assessment.

OBJECTIVES: To determine whether quantitative parameters from contrast-enhanced ultrasound examination (CE-US) of the bowel wall correlate with endoscopic inflammatory activity in Crohn's disease and to assess its utility in the follow-up of patients on treatment. METHODS: This was a prospective study of 30 patients with Crohn's disease requiring colonoscopy or flexible sigmoidoscopy. The Crohn's disease activity index (CDAI), serum C-reactive protein (CRP) and Crohn's disease endoscopic index of severity (CDEIS) were calculated. CE-US was performed within 7 days of endoscopy using intravenous perflutren lipid microsphere and repeated a median of 45.5 days following treatment. Time-intensity curves of bowel wall enhancement were generated to derive the area under the curve (AUC), time to peak intensity (TTP) and peak intensity (PI). RESULTS: In separate multivariate regression models, the AUC, TTP and PI showed no significant association with endoscopic activity whilst controlling for CDAI and CRP. However, the reductions in CDAI and CRP following treatment were mirrored by significant reductions in TTP (P = 0.05). CONCLUSIONS: CE-US does not appear to predict for Crohn's disease endoscopic activity independent of simple parameters such as CDAI and CRP. However, it may be useful in the serial assessment of patients on treatment.

Systemic treatments and outcomes in CIC-rearranged Sarcoma: A national multi-centre clinicopathological series and literature review.

CIC-rearranged sarcoma is a recently established, ultra-rare, molecularly defined sarcoma subtype. We aimed to further characterise clinical features of CIC-rearranged sarcomas and explore clinical management including systemic treatments and outcomes. METHODS: A multi-centre retrospective cohort study of patients diagnosed between 2014-2019. RESULTS: Eighteen patients were identified. The median age was 27 years (range 13-56), 10 patients were male (56%), 11 patients (61%) had localised disease and 7 patients had advanced (metastatic or unresectable) disease at diagnosis. Of 11 patients with localised disease at diagnosis, median overall survival (OS) was 40.6 months and the 1-, 2- and 5-year OS estimates were 82%, 64% and 34% respectively. Nine patients (82%) underwent surgery (all had R0 resections), 8 (73%) patients received radiotherapy to the primary site (median dose 57Gy in 28 fractions), and 8 (73%) patients received chemotherapy (predominantly Ewing-based regimens). Metastases developed in 55% with a median time to recurrence of 10.5 months. In patients with advanced disease at diagnosis, median OS was 12.6 months (95% CI 5.1-20.1), 1-year OS was 57%. Median progression-free survival was 5.8 months (95% CI 4.5-7.2). Durable systemic therapy responses occurred infrequently with a median duration of systemic treatment response of 2.1 months. One durable complete response of metastatic disease to VDC/IE chemotherapy was seen. Responses to pazopanib (n = 1) and pembrolizumab (n = 1) were not seen. CONCLUSION: In this series, CIC-rearranged sarcomas affected young adults and had a high incidence of presenting with, or developing, metastatic disease. The prognosis overall was poor. In advanced disease, durable systemic therapy responses were infrequent.

Localization of cortico-peripheral coherence with electroencephalography.

BACKGROUND: The analysis of coherent networks from continuous recordings of neural activity with functional MRI or magnetoencephalography has provided important new insights into brain physiology and pathology. Here we assess whether valid localizations of coherent cortical networks can also be obtained from high-resolution electroencephalography (EEG) recordings. METHODS: EEG was recorded from healthy subjects and from patients with ischemic brain lesions during a tonic hand muscle contraction task and during continuous visual stimulation with an alternating checkerboard. These tasks induce oscillations in the primary hand motor area or in the primary visual cortex, respectively, which are coherent with extracerebral signals (hand muscle electromyogram or visual stimulation frequency). Cortical oscillations were reconstructed with different inverse solutions and the coherence between oscillations at each cortical voxel and the extracerebral signals was calculated. Moreover, simulations of coherent point sources were performed. RESULTS: Cortico-muscular coherence was correctly localized to the primary hand motor area and the steady-state visual evoked potentials to the primary visual cortex in all subjects and patients. Sophisticated head models tended to yield better localization accuracy than a single sphere model. A Minimum Variance Beamformer (MVBF) provided more accurate and focal localizations of simulated point sources than an L2 Minimum Norm (MN) inverse solution. In the real datasets, the MN maps had less localization error but were less focal than MVBF maps. CONCLUSIONS: EEG can localize coherent cortical networks with sufficient accuracy.

NKX3.1 immunohistochemistry is highly specific for the diagnosis of mesenchymal chondrosarcomas: experience in the Australian population.

Mesenchymal chondrosarcoma (MC) is a rare sarcoma that typically arises in adolescents and young adults and characteristically harbours a HEY1-NCOA2 gene fusion. A recent study has shown that NKX3.1 immunohistochemistry (IHC) is highly specific and sensitive in MCs. NKX3.1 is a nuclear marker expressed in prostatic tissue and is widely used in most laboratories to determine prostatic origin of metastatic tumours. In the current study we investigated whether this stain can be used in the diagnostic workup of MC, as it may assist in triaging cases for further molecular testing, by assessing its expression in a cohort of MCs and in a wide spectrum of sarcoma types. Furthermore, we aimed to elucidate if expression of NKX3.1 by MCs is related to androgen receptor (AR) expression. We identified NKX3.1 positive nuclear staining in 9 of 12 individual patients of MC (n=20 of 25 samples when taking into account separate episodes). Four of the five negative specimens had been previously subjected to acid-based decalcification. NKX3.1 was negative in 536 samples from 16 non-MC sarcomas derived from largely tissue microarrays (TMAs). Overall, we identified 80% sensitivity and 100% specificity for NKX3.1 IHC in MCs. The sensitivity increased to 95.2% when acid-based decalcified specimens were excluded from the analysis. No correlation between NKX3.1 expression and AR IHC was identified. In summary, our findings indicate that NKX3.1 nuclear positivity is highly sensitive and specific for MC, provided that ethylenediaminetetraacetic acid (EDTA)-based rather than acid-based decalcification is used for sample processing. NKX3.1 IHC in the right clinical and histopathological setting can potentially be sufficient for the diagnosis of MC, reserving molecular confirmation only for equivocal cases.

Lateralization of interimplant timing and level differences in children who use bilateral cochlear implants.

OBJECTIVES: Interaural level differences (ILD) and interaural timing differences (ITD) are important cues for locating sounds in space. Adult bilateral cochlear implant (CI) users use ILDs more effectively than ITDs. Few studies investigated the ability of children who use bilateral CIs to make use of these binaural cues. Our working hypothesis was that children using bilateral CIs are able to perceive changes in ITDs and ILDs similar to their normal-hearing (NH) peers. DESIGN: Participants were two groups of children; 19 bilateral implant users (CI) and nine NH children. The children in the CI group had received a second CI after 4.9 +/- 2.8 yrs of unilateral use. Children performed a four alternative forced-choice lateralization task in which they were asked to describe stimuli as coming from the left side, right side, middle of the head, or from both sides simultaneously. Stimuli were 500 msec trains of electrical pulses delivered to apical electrode no. 18 (CI group) or clicks (NH group) presented 11 times per second with either ITDs (0, 400, 1000, or 2000 microsec delay between sides) or level differences (0, 10, or 20 Current Units (CI group) or 0, 10, or 20 dB (NH group) difference between sides). ITDs were presented using current levels that were balanced using left and right electrically evoked brain stem responses. Stimulus levels evoking response amplitudes that were most similar were used. RESULTS: Responses from children in the CI group changed significantly with changes in ILD of bilateral stimuli, but not with changes in ITD. Responses from children in the CI group were significantly different from those in the NH group in three ways. Children in the CI group perceived bilaterally presented electrical pulses: (1) to come from the second implanted side more often than the first, (2) to rarely come from the middle, and (3) to come from both sides of the head simultaneously. Perceived changes in lateralization with ILD changes were correlated with differences in amplitudes of electrically evoked brain stem responses by the left versus right CI. CONCLUSIONS: The results of this study illustrate that children who use bilateral CIs can lateralize stimuli on the basis of level cues, but have difficulty interpreting interimplant timing differences. Perceived lateralization of bilaterally presented stimuli to the second implanted side in many of the stimulus conditions may relate to the use of different device generations between sides. Further differences from normal lateralization responses could be due to abnormal binaural processing, possibly resulting from a period of unilateral hearing before the provision of a second implant or due to insufficiently matched interimplant stimuli. It may be possible to use objective measures such as electrically evoked auditory brain stem responses wave eV amplitudes to provide balanced levels of bilateral stimulation in children who have had no binaural hearing experience.

NTRK-rearranged mesenchymal tumours: diagnostic challenges, morphological patterns and proposed testing algorithm.

Oncogenic fusions involving neurotrophic receptor tyrosine kinase (NTRK) genes are being increasingly identified in a range of mesenchymal tumours unrelated to infantile fibrosarcoma or cellular congenital mesoblastic nephroma, where the canonical ETV6-NTRK3 fusion was first described more than two decades ago. Recognition of these NTRK-rearranged tumours poses a diagnostic challenge to surgical pathologists due to their non-specific clinical and pathological features. However, their recognition is of heightened importance, particularly in locally advanced and metastatic sarcomas, due to the recent availability of selective and highly effective targeted therapy. Herein, we present an Australian multi-institutional series of six of these rare NTRK-rearranged mesenchymal neoplasms to share the local experience and diagnostic challenges as well as to highlight key morphological patterns and immunoprofiles that provide the most helpful clues in routine practice. We also propose a diagnostic algorithm for the detection of these fusions, drawing attention to the limitations of ancillary studies including immunohistochemistry against tropomyosin receptor kinase (TRK) protein, fluorescence in situ hybridisation (FISH) and next generation sequencing.

Clear cell sarcoma of the gastrointestinal tract and malignant gastrointestinal neuroectodermal tumour: distinct or related entities? A review.

Clear cell sarcoma is an uncommon sarcoma which rarely occurs as a primary tumour in the gastrointestinal tract (CCS-GIT). It shares common molecular genetic abnormalities with the more recently described entity, malignant gastrointestinal neuroectodermal tumour (GNET) but is distinguished by its morphological and immunohistochemical findings. The exact nosological relationship between these tumours continues to be debated. In this review, we present two cases of these rare neoplasms from our files and perform a statistical comparison of all published cases to determine if significant differences exist in their clinicopathological features and biological behaviour. Thirteen cases of CCS-GIT and 58 of GNET were included. CCS-GIT occurred more commonly in males (84.6% vs 46.6%, p = 0.01) and in an older age group (median 57 vs 33 years, p < 0.01). There was no significant difference in their location in the gastrointestinal tract, median tumour size and proportion of cases with an EWSR1-ATF1 vs EWSR1-CREB1 fusion. Median survival for CCS-GIT was 13.5 months and for GNET, 9.5 months (p = 0.78). There was no significant difference in the Kaplan-Meier survival curves for either time to first metastasis (p = 0.88) or overall survival (p = 0.18), including after controlling for tumour size using regression models. Our analysis confirms that aside from morphological variations between these tumours, they also exhibit epidemiological and clinical differences. Despite the prevalent perception that GNET is associated with a more aggressive clinical course, our findings indicate that there is no significant difference in their biological behaviour, although both clearly share a bleak prognosis. Further experience is awaited to determine optimal treatment strategies and whether CCS-GIT and GNET would differ in their response to various therapies.

A Comparison of Regularization Methods in Forward and Backward Models for Auditory Attention Decoding.

The decoding of selective auditory attention from noninvasive electroencephalogram (EEG) data is of interest in brain computer interface and auditory perception research. The current state-of-the-art approaches for decoding the attentional selection of listeners are based on linear mappings between features of sound streams and EEG responses (forward model), or vice versa (backward model). It has been shown that when the envelope of attended speech and EEG responses are used to derive such mapping functions, the model estimates can be used to discriminate between attended and unattended talkers. However, the predictive/reconstructive performance of the models is dependent on how the model parameters are estimated. There exist a number of model estimation methods that have been published, along with a variety of datasets. It is currently unclear if any of these methods perform better than others, as they have not yet been compared side by side on a single standardized dataset in a controlled fashion. Here, we present a comparative study of the ability of different estimation methods to classify attended speakers from multi-channel EEG data. The performance of the model estimation methods is evaluated using different performance metrics on a set of labeled EEG data from 18 subjects listening to mixtures of two speech streams. We find that when forward models predict the EEG from the attended audio, regularized models do not improve regression or classification accuracies. When backward models decode the attended speech from the EEG, regularization provides higher regression and classification accuracies.