DESCRIPTION

Field of the Invention

The present invention refers to the field of medical devices, in particular to a system for monitoring a motor disorder, in a subject in need thereof, of the type comprising one or more implantable micro-electrodes inserted in a location in the basal ganglia of the subject and configured for recording data of neural activity of said location, and a computational unit comprising at least a processor configured to carry out monitoring steps of the neural activity.

State of the Art

Deep Brain Stimulation (DBS) is a well-known technique for treating a number of neurological disorders, and especially motor disorders. To mention only the most common motor disorders, an ever-growing population of millions of people are affected by Parkinson’s Disease. Most patients eventually reach a stage in which the effect of pharmaceutical therapies in limiting motor symptoms is limited and in these cases DBS becomes an effective way to improve the patient conditions. In this technique, the localization of the optimal stimulation site and the determination of the optimal stimulation pattern is a key aspect, of paramount importance to ensure the efficacy of the therapy and to avoid collateral effects due to stimulation of areas not involved in the disorders.

Indeed, several techniques address the issue of finding the right nucleus within the basal ganglia - which typically is the Subthalamic Nucleus for Parkinson’s Disease. However recent studies highlighted the fact that the localization of the stimulation within the selected area determines the probability of collateral effects (in particular cognitive disorders) and ultimately the efficacy of the therapy. This also applies to other movement disorders as for instance the localization of the optimal stimulation within the Globus Pallidus Internus to suppress tics originated by Tourette’s Syndrome.

Pre-surgery imaging does not provide a sufficient resolution and visual inspection of recordings does not provide the necessary information. In the disclosure of patent publication W02004093653A2 there is proposed to use arrays of features extracted from micro-electrode recordings (referred to also as “microrecordings”) to identify the transition of the nuclei and to classify them. These features rely on power spectral density and differences in spike count, but the disclosed system and the parameters used have proven themselves unsatisfactory in connection with the evaluation of specific locations within a target area and more generally speaking with the accurateness of the evaluation and possibility to obtain meaningful information and instructions to be successfully used in the DBS treatment.

Another disclosure in document US20100204748A1 proposes to analyse microelectrode recordings to identify the oscillatory dorsolateral region of the subthalamic nucleus characterized by the presence of oscillatory neurons in the beta range. This method, optimized only for Parkinson’s Disease, has substantially the same limit of the one mentioned just above.

In yet another disclosure, the system of document US7974696 provides for evaluating the effectiveness of a certain stimulus applied through intracranial microelectrodes inserted into the basal ganglia. For the position, fixed, at which the stimulus is applied, feedback is obtained aimed at improving the type of stimulus, in particular by seeking a variable that correlates over time with the subject's motor state (onset/offset of a symptom) such that the amplitude of stimulation is adjusted online. This disclosure makes no contribution to the aspect of finding more suitable positions for permanent DBS implantation.

There is therefore a strongly felt need for a system capable of monitoring the motor disorder, in the sense that the disorder is evaluated as it affects various possible locations, and consequently the specific locations in the basal ganglia are assessed as a possibly favourable site for the DBS implant.

Summary of the Invention

The object of the present invention is therefore to solve the problems mentioned above by providing a system capable to monitor the neural activity of the basal ganglia of a subject with the aim of successfully assessing the specific affection of a certain disorder on a specific location and consequently classifying the location in its appropriateness to be the optimized location for a DBS implant.

A particular object of the present invention is to provide a system of the above- mentioned type, which is also useful for setting up disorder-related and location-related control instructions for the DBS electrical stimulation to be carried out at the location.

A further object of the present invention is to provide a computer program product comprising processor-executable instructions for causing the operation above mentioned system.

Still a further object of the present invention is to provide a computer-readable storage medium having stored thereon the above said computer program.

These and other objects of the present invention are achieved by a system for monitoring a motor disorder, in a subject in need thereof, comprising: - one or more implantable micro-electrodes previously inserted in at least two distinct locations in the basal ganglia of said subject and configured for recording data of neural activity of said locations; - a computational unit comprising at least a processor configured to: -- acquire the data of neural activity from the one or more micro-electrodes; - decode and record from the acquired data at least one feature that correlates with a motor state of said subject, wherein the at least one decoded feature is selected from among one or more of: at least one feature linked with spiking temporal pattern of putative single-unity activity; at least one feature linked with spectral component of background unit activity; - run a disorder-related comparison function between said at least one decoded feature and a pre-loaded and disorder-related set of reference data corresponding to an abnormal motor state of said subject, outputting a DBS efficacy classification of said at least two locations based on the different response of said function between said at least two locations.

In an embodiment, said processor is configured to decode at least one feature linked with the spiking temporal pattern selected from among one or more of: a regularity index of the spiking; spiking shape factor; spiking firing rate. The regularity index of the spiking may comprise e.g. a ratio between bursting activity and tonic activity.

In another embodiment the processor is configured to record both the spiking temporal pattern of putative single-unity activity and the spectral component of background unit activity, and to use a combination of the one or more relevant features for comparison with said reference data.

The processor may be further configured to output location-sensitive DBS control instructions, said instructions being appropriate for suppressing symptoms of the disorder associated with said reference data, and can further comprise a DBS electrical stimulation module adapted to be implanted in said locations and configured to receive the instructions from said processor.

The comparison function and the reference data are preferably related with a disorder selected from among one of: Parkinson’s disease; Tourette’s syndrome; dystonia; essential tremor.

The pre-loaded reference data, and/or the comparison function and/or, when applicable, the DBS control instructions, can be obtained based on clinical data of previous patients.

In another aspect of the present invention, there is provided a computer program product comprising processor-executable instructions to cause a processor in a computational unit of a system as defined above, to execute the steps of: -- acquiring the data of neural activity from the one or more micro-electrodes; -- decoding and recording from the acquired data at least one feature that correlates with a motor state of the subject, wherein said at least one decoded feature is selected from among one or more of: at least one feature linked with spiking temporal pattern of putative singleunity activity; at least one feature linked with spectral component of background unit activity; - running a disorder-related comparison function between said at least one decoded feature and a pre-loaded and disorder-related set of reference data corresponding to an abnormal motor state of said subject; -- outputting a DBS efficacy classification of said at least two locations based on the different response of said function between said at least two locations.

The invention also provides a computer-readable medium having stored thereon a computer program as mentioned above.

Brief description of the drawings The features and advantages of the system according to the invention will be more clearly illustrated in the following exemplary, non-limiting description of the embodiments thereof with reference to the attached drawings, wherein:

-Figure 1 is a schematic block-diagram functional representation of the configuration of the system according to the present invention;

-Figures 2a, 2b are representations in connection with an embodiment of the system operating with the feature of spiking activity signal decoded from the neural activity at locations in the basal ganglia of a subject;

- Figures 3a, 3b are representations in connection with an embodiment of the system operating with the feature of spectral component analysis of the background unit activity at locations in the basal ganglia of a subject; and

- Figure 4 is a diagram showing the shape factor as a function of beta band power in recordings conducted with the system at various locations.

Detailed description of the invention

The invention provides a system capable of analysing the temporal structure of the spiking activity acquired with microelectrode recordings (MER), to be used in the context of DBS implant surgery, to evaluate the target nucleus as an appropriate implant site, and determining the optimal stimulation location and stimulation frequency.

The inventors have focused their consideration on the activity of neurons within each nucleus of the basal ganglia, noting that this activity is far from being spatially homogeneous, hence neuronal dynamics can provide a map of the functional structure of the nucleus. In particular, pathological activity might be localized to specific subregions of the nuclei, coherently with the hypothesis that each subregion maps specific functions/cortical areas, hence neuronal dynamics can even localize specific targets to affect selectively the neural circuits displaying pathological activity and iii) the analysis of the temporal structure of the pathological neural activity can be used to characterize the disorder to estimate the optimal stimulation frequency.

Accordingly, use is proposed, as neural markers, of a set of features of the temporal structure of the activity recorded through MER (Micro Electrode Recording) at locations in the volume of the patient’s basal ganglia. These features include spiking regularity of single-unit activity and spectral analysis of the high frequency component of the background activity in the MER, and are used to classify locations e.g. at different inspected depth, determining if they are appropriate for DBS implant and following stimulation, and to also to propose a suitable stimulation frequency

As shown in the exemplifying scheme of Figure 1 , the main input to the system is the raw micro-electrode recordings. First of all the spiking activity is separated from the background with thresholding and if necessary sorting of the different units. At least one of the two signals is then analysed, but possibly both are analysed separately in a parallel way.

Spiking activity can be typically analysed as a discrete series of events. Possible features, that - as such - can be derived from known signal processing techniques, include i) spike count, ii) difference from re-aligned templates, iii) bursting index, iv) intra- and inter- burst frequency, and v) regularity measures, convolution with kernels (e.g. exponential kernels) and vi) spectral analysis. These features are explained in detail in the following: i) Spike count can be performed over a reference window and compared with reference value of optimal depth. If optimal depth is characterized by a specific temporal structure rather than a specific intensity, the quality of each depth can be measured in several ways: ii) by extracting an average template and then compare the template with the recorded spike train by means of principal component analysis or Victor-Purpura distance, with a method introduced in Oddo et al., 2017; or by identifying bursting activity with the ranking surprise or other methods and then measuring; iii) the fraction of spikes fired during bursts, or the total time spent bursting, iv) or the average time interval between bursts or the average firing rate within bursts; v) or evaluating the overall irregularity of the recordings by fitting the inter-spike interval distribution with a gamma function and then using its shape factor as measure of irregularity (Vissani et al., 2019). All the aforementioned features operate in the time of the discrete events. By convoluting the spikes with kernels (e.g., exponential kernels) or by analysing the background activity the quality of the inspected depth can instead be measured by spectral analysis which means; vi) identifying the frequency and the amplitude of a prominent spectral peak, or measuring the power within a pre-defined band (e.g beta band [13 30] Hz), or finally and measure the coherence between the oscillations over different bands and the spiking activity.

The parameters obtained from these analyses are then compared to corresponding pre-loaded reference set of data, chosen based on the disorder to be treated and expressive of an abnormal motor state of the subject. As a result of said comparison, providing different responses between the different positions at which the microrecordings were taken, a classification/ranking of the location is outputted in terms of its efficacy as a site for DBS stimulation.

As a more detailed example, Figure 2 shows a possible procedure based on the shape factor of the inter-spike interval distribution (Vissani et al., 2019). The starting step is the finding of a significant relationship between the distance from the optimal target location, defined as a location that reliably induced an improvement in the clinical scales (for instance, YTGSS for Tourette or UPDRS for Parkinson) and a microrecording features as the shape factor. For instance, microrecordings in the proximity of the optimal target can have a particular low shape factor (indicating high irregularity) as depicted in Figure 2a. Then, during a novel DBS implant in which the optimal target has to be determined, microrecordings are acquired at different depths within the target area (e.g., the subthalamic nucleus) and the shape factor of each recording is determined after few seconds. In Figure 2B two different locations are considered. The shape factor at the Location 2, as compared with reference data related with an efficient stimulation suppressing abnormal motor condition of the subject, is classified as a preferred or possibly optimal location whereas e.g. Location 1 is categorized, based on analogous comparison considerations, as tendentially unsuitable, or less suitable.

A second example relies instead in the analysis of the background activity, i.e., the continuous signal acquired through microrecordings low pass filtered to remove action potentials with spectral analysis tools (again, known as such), for instance measuring the spectral content across different frequency bands it is possible to find out that close to the optimal target location (defined as above) there is an excess beta band power (see Figure 3a). Then, during a novel DBS implant in which the optimal target has to be determined, microrecordings are acquired at different depths within the target area (e.g., the subthalamic nucleus) and the beta power is computed online over windows of few seconds. The power spectrum of location 1 displays a stronger beta peak and hence an overall excess beta power. Based on a similar approach as described in connection with the spiking temporal pattern, Location 1 is here associated to a higher efficacy classification.

The classification step can be performed by making use of a single feature, or with a weighted linear combination of them, or by clustering their values with standard techniques, e.g., support vector machine algorithms. It is important to notice that all the aforementioned quantities can be extracted by a single recording at each depth, so they can be extracted and processed in parallel and then combined to predict optimal location. In an example combining the previous two examples, both the analysis depicted in Figure 2a and the one depicted in Figure 3a hold: close to target location shape factor is low and beta power is high (two signs of particularly strong beta bursts). The optimal target can then be selected as a location in which the microrecording displayed at the same time low shape factor (Figure 4).

Based on the characteristic timescales of pathological oscillations (for instance inter-burst spiking intervals and/or background beta frequency peak), and/or the spatial distribution of pathological oscillations, the system can be programmed so as to output instructions for the DBS module to deliver optimal stimulation frequency to suppress the symptoms of the disorder to be treated.

By way of further example, the target nucleus of the location can be the subthalamic nucleus, and if the pathology/disorder to be treated is Tourette’s Syndrome, the feature considered can be the shape factor of the spiking pattern. For dystonia disorders, the target nucleus can be the Globus Pallidus internus, and the feature considered by the system for its assessment can be the firing rate.

In contrast to known solutions such as that of US7974696, according to the invention a position in which to permanently implant the DBS is sought, and this is to be established within a limited time frame in which typically the patient will not exhibit any transient symptoms. Then the correlation with the presence of the symptoms from which the patient suffes is compared at various locations in space (not time), basically looking for the location of the subcircuit most associated with the symptom.

The system is configured to carry out the same steps for each disorder for which DBS is recommended. The system is set up with a specific disorder-related configuration by selecting the specific features performing the classification, and the reference data and comparison function/criterion for the feature, features or combination thereof. The reference setup can be inserted manually by the user or downloaded from a website associated to the system. The reference set up, as well as the possible instructions on the control of the DBS stimulation module, can be determined with training on acquired datasets, e.g., by maximizing the mutual information between each feature and the location.

Variations and/or modifications can be brought to the system according to the invention without departing from the scope thereof as defined by the attached claims.

Bibliography

Oddo CM, Mazzoni A, Spanne A, Enander JMD, Mogensen H, Bengtsson F, Camboni D, Micera S, Jorntell H (2017) Artificial spatiotemporal touch inputs reveal complementary decoding in neocortical neurons. Sci Rep 7:45898.

Vissani M, Cordelia R, Micera S, Eleopra R, Romito LM, Mazzoni A (2019)

Spatio-temporal structure of single neuron subthalamic activity identifies DBS target for anesthetized Tourette syndrome patients. J Neural Eng 16:066011.